.. .. .3.
Eu... n... .

v:

 

-- . .

. . . . . . l
. . . . . r x
. . . . n.n.fin{..m.wu
. . . . .2 .

. .. m... in -.

. €51.39. .

. daqvtx 2 1

. ..

.
1w... .
s .fimflwnnm. ,
3......” .
.mmmwrrm . . . .
..._ . . . .. . . . 1. .....,...._m..... ..
. . . . 4:71.: u.
. . . ”mfivw....mwnw.b:‘h

.

75.
($1 .19...
1.

hr!» . ~
afimwxma...
«.mMam. 1
. .

z .
3.

1

 

 

 

 

8r
.. )
s. E. ii.

 

nu... -193... .

~Jl’u’.f.~ h d-lnal.

.... -1-.- 1
a... 1.... £5“...
r 1... 1.... .1...
.11 1.1.9:).

 

. .$.1....... . . . .. . . . I!
imamfidm. 11.11.... .1 .3 . . . . $523. ,3 in. .1... N..-.....m....Hm.

"IIII!

 

UNIVERSITY LI I8|RARIES

lllllllllllllllllllllllll ||llll|33l4| | 44...

? 31293 017141
.2

 

 

 

 

 

 

 

 

 

 

 

 

 

 

This is to certify that the

thesis entitled

MATRIX METALLOPROTEINASES IN TISSUE REMODELING,
REGENERATION, AND PROSTATE CANCER

presented by
William Winston Chu

has been accepted towards fulfillment
of the requirements for

M.S. Zoology

degree in

 

 

Dr. MukEa M. Webber

 

Major professor

Date w °

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

 

“LIBRARY

MIchIgan State
Unlvorelty

 

 

 

_ PLACE IN RETURN BOX
to remove this checkout from your record.
TO AVOID FINES return on or before date due.

 

MTE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1193 MM“

 

MATRIX METALLOPROTEINASES IN TISSUE REMODELING, REGENERATION
AND PROSTATE CANCER

. By

William Winston Chu

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Zoology

1998

ABSTRACT

MATRIX METALLOPROTEINASES IN TISSUE
REMODELING REGENERATION AND PROSTATE CANCER

By

William Winston Chu

The turnover of extracellular matrix (ECM) occurs during tissue remodeling,
regeneration and cancer. Matrix metalloproteinases (MMPs) play a key role in ECM turnover
because they have the ability to degrade the ECM, especially collagen, a major component of
the extracellular matrix. MMPs are, therefore, involved in ECM turnover during normal
development and wound healing as well as arthritis and cancer. The activity of MMPs is
highly regulated at three levels: 1) synthesis and secretion; 2) activation and 3) extracellular
proteolytic activity. The extracellular protease activity is regulated by specific, tissue inhibitors
of matrix metalloproteinases or TIMPs. This tight regulation allows for the spatial temporal
control of MMP activity required under normal conditions where homeostasis in ECM
turnover is maintained by on a balance between active MMP and TIMP. A loss in this balance
leads to pathological conditions. For example, a proportional increase in TIMP levels results
in accumulation of ECM as seen in glomerular and arterial sclerosis. Similarly, a proportional
increase in MMP levels results in increased degradation of ECM, as seen in arthritis and
cancer. This study reviews the family of MMPs, their substrates and inhibitors, and examines

their roles in both normal and pathological conditions.

ACKNOWLEDGEMENTS

There are several people who have helped me along my journey toward the completion
of this thesis. I convey my deepest gratitude to my mentor, Dr. Mukta M. Webber who has
been my teacher, motivator, sounding board, editor, confidant and friend. I am very thankful
for all that she has given me over the years and I hope that this thesis is the first of many of
my works, which will reflect her influence. I thank my committee members Dr. Stephen C.
Bromley, and Dr. Daniel E. Williams for their participation in this project as well as their
advice and support. I thank my colleagues here at M.S.U., Diana Bello-DeOCampo, Salmaan
Quader, Nicholas Tralcul, Rebekah Sharp, Anthony Zander, Robert D. Hollister, Ian Ramj ohn,
Skip J. Van Bloem, and Stefanie L. Whitmire not only for their assistance in the completion
of this thesis but also their friendship.

In addition, I would like to express my sincere appreciation to Michigan State
University; the Collage of Natural Science, in particular Dr. Estelle McGroarty, and Dr. Pui
Kei Wong; the Department of Zoology, especially Dr. Thomas Burton, and Dr. Richard Hill
for their council and support. I thank Dr. GustaafA de Zoeten and the Depanment of Botany and
Plant Pathology for allowing me to use their computer facilites.

Finally, I wish to thank my family, William Chu, James and Cynthia Shea, C. Michael
O’Shea, Kelly M. Shea, Fredrick and Kimberly Hinz, Karina E.E. Hinz, and Karalee Hinz,

for their everlasting love and support.

TABLE OF CONTENTS

 

ABSTRACT ..................................................................................................................... ii
ACKNOWLEDGMENTS ................ - ................... iii
TABLE OF CONTENTS ......................................................................................... iv
LIST OF TABLES ..................................................................................................... xii
LIST OF FIGURES ................................................................................................. xiii
LIST OF ABBREVIATIONS .............................................................................. xx
CHAPTER 1
COLLAGEN CLASSIFICATION ................................................... 1
ABSTRACT .................................................................................................. 1
INTRODUCTION ......................................................................................... 1
STRUCTURE OF THE COLLAGEN MOLECULE ..................................... 3
Collagen triple helix ............................................................................ 3
Nomenclature for a-chains ................................................................ 5
Collagen non-helical domains ............................................................ 6
CLASSIFICATION OF COLLAGENS ................................................................ 6
Fibril-forming collagens ....................................................................... 9
Collagen in cartilaginous tissues ...................................................... 9
Collagens in non-catrilaginous tissues .............................................. 10
Non-fibril Forming collagens .............................................................. 12

iv

COLLAGEN BIOSYNTHESIS .......................................................................... 16

The procollagen model ....................................................................... 16
Intracellular events ............................................................................ 20
Collagen Genes ...................................................................... 20

Translational, co-translational,

and early post-translational events .............................. 24
Extracellular events ........................................................................... 28
Microfibrilsform quarter staggered arrays .............................. 28
Aggregation ........................................................................... 28
CONCLUSIONS .................................................................................................. 30
REFERENCES .................................................................................................. 3 1
CHAPTER 2
MATRIX METALLOPROTEINASES (MNIPs):
CLASSIFCATION AND
BIOCHEMICAL PROPERTIES .................................................................... 33
ABSTRACT ....................................................................................................... 33
INTRODUCTION ............................................................................................. 33
CLASSIFICATION OF
MATRIX METALLOPROTEINASES ............................................................. 37
Collagenases ..................................................................................... 3 8
Gelatinases ........................................................................................ 38
Stromelysin ....................................................................................... 40
Other MMPs ........................................................................................ 40

BIOCHEMICAL PROPERTIES OF MMP ...................................................... 43

Molecular weight characteristics ...................................................... 43
Compositional analysis .................................................................... 43
MMP domain structure ..................................................................... 44
Metal ion requirement ..................................................................... 47
MODULATION OF MMP ACTIVITY ............................................................ 50
CONCLUSIONS .................................................................................................... 50
REFERENCES .................................................................................................. 5 1
CHAPTER 3
MODULATION OF
MATRIX METALLOPROTEINASES ....................................................... 55
ABSTRACT ...................................................................................................... 55
INTRODUCTION ............................................................................................ 55
EXPRESSION OF MMPs ............................................................................... 56
ACTIVATION OF MMPs ................................................................................ 62
The cysteine switch ......................................................................... 63
INHIBITION OF THE ACTIVATED MMP ................................................... 64
ROLES FOR MATRIX
METALLOPROTEINASES ............................................................................. 67
CONCLUSIONS ................................................................................................. 69
REFERENCES .................................................................................................... 70

CHAPTER 4
TISSUE INHIBITORS OF MNIPS (TIMPS):
STRUCTURE, EXPRESSION,

GENETIC REGULATION, AND FUNCTION ........................................ 73
ABSTRACT ...................................................................................................... 73
INTRODUCTION ............................................................................................ 73
THE TIMP FAMILY ......................................................................................... 74
DISTINCT SECONDARY SITES OF
TIMP I MMP INTERACTION ........................................................................ 76
TIMP-3 IS BOUND TO THE ECM ................................................................ 8O
PREFERENTIAL INHIBITION OF MMP BY TIMP .................................. 82
DIFFERENTIAL EXPRESSION OF TIMP ................................................... 83
REGULATION OF THE TIMP GENES ......................................................... 83
TIMP EFFECTS ON CELL GROWTH .......................................................... 87
APPLICATIONS ................................................................................................ 90
CONCLUSIONS .................................................................................................. 92
REFERENCES .............................................................................................. 93

E:

CHAPTER 5
THE ROLE OF MATRIX METALLOPROTEINASES

IN INVASION BY CANCER CELLS .......................................................... 97
ABSTRACT ...................................................................................................... 97
INTRODUCTION ........................................................................................... 97
DISTRIBUTION OF MMPS INVOLVED IN
BASEMENT MEMBRANE INVASION ...................................................... 100

Gelatinase A ................................................................................... 100
Gelatinase B ............................................................................... 101
Stromelysins .................................................................................. 101
Matrilysin ..................................................................................... 102
MT-MMPs ................................................................................... 102
EXPRESSION OF MMP GENES ................................................................. 103
MECHANISMS OF MMP ACTIVATION ................................................... 105
Activation of MMPs by proteolytic enzymes .................................. 105
Autoactivation of MMPs .................................................................. 107
Activation of MMP by furin ............................................................. 107
Mechanism for the activation of progelatinase A ........................ 108

EVIDENCE FOR THE ROLE OF MMP IN INVASION

THOROUGH THE BASEMENT MEMBRANE ........................................... l 10
CONCLUSIONS ................................................................................................ 1 14
REFERENCES .............................................................................................. 1 15

E;

CHAPTER 6
MATRIX MET ALLOPROTEINASES

AND PROSTATE CANCER ............................................................................ 118
ABSTRACT ...................................................................................................... 118
INTRODUCTION .......................................................................................... 118
TUMOR CELL INVASION OCCURS INTHREE STEPS ........................... 121
EXPRESSION OF MMP IN PROSTATE CANCER ..................................... 121
MODULATION OF MMP ACTIVITY IN PROSTATE CANCER ............. 126
LOSS OF INHIBITION BY TGF-B ................................................................ 128
MMP SECRETION AND INVASION ........................................................... 129
APPLICATIONS .............................................................................................. 131
CONCLUSIONS ................................................................................................ 137
REFERENCES .............................................................................................. 138

CHAPTER 7

COMMON MECHANISMS OF
MATRIX METALLOPROTEINASES
IN NORMAL AND PATHOLOGICAL

CONDITIONS INVOLVING
EXTRACELLULAR MATRIX TURNOVER .......................................... 140
ABSTRACT ........................................................................................................... 140
INTRODUCTION ................................................................................................. 140
NORMAL CONDITIONS .................................................................................. 141
Development ...................................................................................... 141
Endometrial cycling ......................................................................... 1 51
Regulation of mammary involution ................................................. 153
Wound healing .................................................................................... 156
Nerve regeneration ........................................................................ 1 5 8
PATHOLOGICAL CONDITIONS ....................................................................... 160
Arthritis ............................................................................................. 160
Cancer ............................................................................................... 162
Glomerular Nephropathy ................................................................... 165
APPLICATIONS ............................................................................................ 167
CONCLUSIONS .................................................................................................... 170
REFERENCES .................................................................................................. 171

CHAPTER 8
SIGNIFICANCE AND APPLICATION

OF MMPs AND TIMPs .................................................................................... 175
SIGNIFICANCE ............................................................................................. 175
ARTHRITIS ..................................................................................................... 176
CANCER ......................................................................................................... 176
CONCLUSIONS ............................................................................................ 177
REFERENCES ................................................................................................. 178

LIST OF TABLES

Table 1.1 . F ibril Forming Collagens and Non-fibril Forming Collagens
Table 1.2. Chain Composition and Distribution for Some Human Collagens
Table 1.3. Hydroxylation Enzymes Involved in Collagen Biosynthesis
Table 2.1 Human Matrix Metalloproteinsaes

Table 2.2 Traditional Classification of Matrix Metalloproteinases

Figure 1.1.

Figure 1.2.

Figure 1.3.

Figure 1.4.

Figure 1.5.
Figure 1.6.

Figure 1.7.

Figure 1.8.

Figure 1.9.

LIST OF FIGURES

The triple helical structure of the collagen molecule. A glycine located every
third residue in the primary structure, fold the molecule so that 8.6 Angstroms
separate each glycine residue. The triple helical turn runs 86 Angstroms. (N imni
1988).

Nomenclature for or-Chains of Type I Collagen

Collagens contain triple helical (solid and open rods) and globular domains
(open and filled circles). Portions of the initially synthesized molecules are
removed prior to their incorporation into insoluble matrices (open rods and circles)
and the rest of the molecule remains intact in the matrix (closed rods and circles).
The domains and their distributions are drawn approximately to scale (N imni
1988)

Type IV collagen schematic diagram showing the short helical domain (7S) in
the C-terminus, the long triple helical domain (TH) and the N-terminus globular
non-collagenous (N Cl) domain (Kucharz 1992).

Proposed models for type IV collagen super structure (Glanville 1987).

The collagen synthesis pathway.

A schematic representation of type I procollagen, showing the “pro” segment,
the helical domain, and the non-helical globular carboxy group (Burgeson 1992).

Diagram of ancestral Collagen Gene and a possible mechanism of collagen
evolution (Kucharz 1992).

Schematic diagram of the human collagen type I alpha 1 chain (Kuhn 1986).

Figure 1.10. The intracellular sequence of events leading to the formation of the procollagen

molecule (Nimni 1988).

Figure 1.11. The hydroxylation of proline in collagen protein (Kucharz 1992).

Figure 1.12. The extracellular aggregation of the collagen molecule and formation of the

native fibril (Burgeson 1992).

Figure 2.1 Tadpole tissue explant digesting collagen. The dark area around the tissue explant

Figure 2.2.

Figure 2.3.

is the area of collagen lysis (Gross 1962).

Domain structure for several MMPs. Pre, leader sequence; Pro, prodomain;
Cat, catalytic domain; H, hinge sequence; HEM, hemopexin-like sequence; F,
furin recognition site; TM, transmembrane domain; FN, fibronectin-like binding
region; C, collagen-like binding region.

Schmatic diagram of conserved domain structure in MMPs (Docherty 1993).

Figure 2.4 Schematic fiinction of the H20 ligand in active zinc site of MMP. S, substrate; B,

Figure 2.5.

Figure 3.1.

Figure 3.2.

Figure 3.3.

Figure 4.1 .

Figure 4.2.

base (Vallee 1992)

Schematic diagram of the “Cysteine Switch”, the interaction of the zinc atom in
the catalytic site, with a cysteine in the prodomain of MMP (Docherty 1992).

Schematic diagram of the promoter of MMP genes, showing the presence of the
AP—l and PEA-3 response elements. The AP-l response element is transactivated
by the dimerization of c-fos and c-jun. The PEA-3 response element is
transactivated by the c-ets family of nuclear transcription factors.

Human fibroblasts were treated with TNF-a for 0, 6, and 24 hours (lanes 1, 2, &
3 respectively) in confluent primary cultures. An increase in the expression of
collagenase is seen with prolonged exposure to TNF-a. TNF-a activates the
AP-l element resulting in an increased expression of collagenase. Col =
collagenase, a—tub = a—tubulin, used as a control (Brenner 1989).

Sequence of the 5’ end of the murine TIMP gene. AP-l is represented by the
open box. All positions are numbered in relation to the transcriptional start site
(Campbell 1991).

The mean serum MMP:TIMP ratio in patients with urothelial cancer. Patients
with recurrence have a significantly higher mean serum MMP:TIMP ratios than
those patients who had no recurrence (Gohji 1996).

Comparison of the structures of the "limp-3 (top row) and Timp-l genes (bottom
row). All features of the Timp-3 gene are indicated in bold type, whereas those
of the Timp-l gene are italicized. The exons are represented by the amino acid
sequences they encode and are enclosed in boxes. Conserved cysteine residues
are in bold type and indicated by solid circles. Arrowheads indicate that the
splice junction splits the codon and that the split is afier the first nucleotide in
this codon. The asterisk indicates the only published splice site for the TIMP2

xiv

Figure 4.3.

Figure 4.4.

Figure 4.5.

gene, this splits the relevant codon of TIMP-2 in an identical fashion. Other
splice junctions are between codons. Potential N-linked glycosylation sites in
TIMP-1 are underlined, while a single site in TIMP-3 is overlined. The sequences
of murine TIMP-3 and TIMP-1 were aligned using the DNASTAR software
with gaps introduced for optimal alignment (Apte 1995).

Protein folding in TIMP-1, showing the conserved cysteines and formation of
loops in TIMP (Caterina 1997).

A schematic representation of the interaction of TIMP-2 and gelatinase A by a
site distinct fi'om the catalytic site of MMP-2.

Comparison between the S-flanking regions of the three TIMP genes. Putative
transcriptional regulatory elements are shown in the promoter of the hTIMP-2
gene and compared with those found in the murine TIMP-1 gene and the murine
TIMP-3 gene. The positions of these elements are determined from the position
of the major transcriptional start site in the three genes. The first two exons in
the murine TIMP-1 gene and the first exon of the hTIMP-2 and murine TIMP-
3 genes are represented by boxes in which the coding part is shown as a black
box with the position of the initiator ATG codon indicated. A 0.8-kb sequence
corresponding to the 5’-end of the human TIMP-l gene contains features similar
to those found in the murine TIMP-1 gene (Hammani 1996).

Figure 4.6. Effect of batimastat (BB-94) treatment on MDA-MB-435 primary tumor regrowth

Figure 5.1.

Figure 5.2.

in athymic nude mice. Tumor growth is measured by tumor volume, treatment
with batismastat decreased tumor growth (Sledge 1995).

Actvation of MMP by plasmin (Corcoran 1996).
The activation of gelatinase A by MT-MMP and TIMP-2. The biniding of

TIMP-2 may be mediated by an unknown integral membrane protein (7)
(Corcoran 1995).

Figure 5.3. Histology of tumors formed by TCLl and SCg6 cells. Hematoxylin/eosin-stained

paraffin sections of primary tumors (panels a and b), invaded skin (panel c) and
lung metastasis (panel d) generated by TCLl (panels a and c) and SCg6 (panels
b and d) cells injected into the mammary gland of nude mice. The spindle cell
tumors (T), entrapped adipose tissue (asterisks) and epithelial ducts (arrowheads)
in the mammary gland are indicated. Comparison of the regulation of MMP-3
expression by ECM in TCLl and SCg6 cells with the nonmalignant, functional
cell line SCp2 revealed striking differences that could play a role in the acquisition
of an invasive tumor phenotype (Lochter 1997).

Figure 5.4. Invasion of Matrigel by SCp2, TCL1, and SCg6 cells and inhibition by proteinase

inhibitors and antisense ODNs. SCp2, TCL1, and SCg6 cells were plated on
Matrigel in modified Boyden chambers. After fixation with 5% glutaraldehyde
and removal of cells that had not migrated, the number of cells that had migrated
through Matrigel after 18 h of culture was quantified by counting the number of
toluidine blue-stained cells per visual field. Panel a, comparison of invasion by
SCp2, TCLl and SCg6 cells. Panel b, TCLl cells (black bars) and SCg6 cells
(white bars) were maintained in the absence of proteinase inhibitors (control) or
with GM6001 (10 nM), aprotinin (0.3 uM), leupeptin (1 11M), or pepstatin (1
uM). Panel c, SCg6 cells were maintained in the presence of GM6001 (white
circles) or GM1210 (black circles) at increasing concentrations given in 11M.
Panel (1, TCLl cells (black bars) and SCg6 cells (white bars) were maintained in
the absence of ODNs (control) or in the presence of SL1 antisense (3.2 nM,
SL1 AS) or sense (3.2 uM, SL1 S) ODNs, SL3 antisense ODNs (8 uM, SL3
AS), or collagenase-3 antisense ODNs (32 uM, C-3 AS) ODNs. Results were
normalized with control values set to 100 in b-d. Error bars indicate standard
deviations from three experiments. (Lochter 1997)

Figure 6.1 Tumor cell invasion occurs in three steps.

Figure 6.2.

Figure 6.3.

Figure 6.4.

Figure 6.5.

Figure 6.6.

Northern blot analysis of gelatinase A mRN A from prostate tumor tissue (T,
lanes 1-4) and adjacent normal stromal tissue (N, lanes 5-8) respectively. In
panel A, the tissue was from benign prostatic hyperplasia (BPH) tissue (lanes 1
and 5): and from prostate tumor tissue with a Gleason score (GS) of 1 (lanes 2
and 6); 2(lanes 3 and 7); or 3(lanes 4 and 8). In panel B the Gleason score was
4 (lanes 1 and 5); 5 (lanes 2 and 6); 6 (lanes 3 and 7); or 8 (lanes 4 and 8)
(Stearns 1993).

Examiniation of gelatinase expression, gelatinase activity and TIMP expression
in PC primary carcinoma of the prostate; NAP, normal adult prostate; and JP,
juvenile prostate. *indicates no significant reading (Lokeshwar 1993).

Immunohistochemical staining of SCID mouse diaphragmatic tumors. DU—145
cells are shown in panel A, and DU-145 cells transfected with matrilysin cDNA
are shown in panel B. Panel B shows cells crossing the basement membrane
(arrows) (Stearns 1993).

Concentration-dependent effect of Bar series (1) IL-10, (2) IL-6, (3) IL-2, and
IL—4 on the production of MMP-1 by HPCA-lOaI-IPV18 prostate cancer cell
line (Wang 1996).

Concentration-dependent effect of Bar series (1) IL-lO, (2) IL-6, (3) IL-2, and
(4) IL-4 on the production of TIMP-1 by HPCA-lOaHPV18 prostate cancer
cell line (Wang 1996).

Figure 6.6. Normal prostate gland showing strong immunoreactivity (indicated by arrows)

Figure 6.7.

Figure 6.8.

Figure 6.9.

Figure 7.1.

Figure 7.2.

Figure 7.3.

for MMP-2 in the basal cells and little immunoreactivity in the secretory epithelium
(Montironi 1996).

High grade prostatic intraepitheial neoplasia shows intense MMP-2 staining
(indicated by arrows) in cells mainly in the cell layer adj cent to the stroma
(Montironi 1996).

Prostatic adenocarcinoma shows intense immunostaining for MMP-2 (arrows)
in small clusters of neoplastic cells located in the periphery of the tumor nodules
at the level of advancing front of the tumor.

Irnmunofluorescence images of PC-3 ML cells labeled with beta-tubulin anti
bodies and a secondary antibodiy conjugated to fluoresein isothiocyanate. Panel
A, show untreated cells. Panel B, shows PC-3 ML cells treated with 0.5 mM
taxol for 6 hour, the more intense signal is due to the stabilization of microtubule
bundles (arrows). This stabilization had an inhibitory efl‘ect on MMP-2 secretion
(Stearns 1992).

Cross-sections of the anterior part of the small intestine hybridized with the
antisense stromelysin probe. The small intestine at stage 59 is shown in panel a.
The layer of connective tissue (CT) is thin except for the typhlosole (Ty). Signals
(arrows) are observed in some cellsof the connective tissue near the muscular
layer (M), but are weaker in the upper region of the typhlosole x150. Panel b
shows a higher magnification of the bottom region of the typhosole at stage 59.
Most of the cells in the connective tissue show a positive reaction x610. The
small intstine at stage 61 is shown in panel c. The basal surface of the epithelial
primordia (astrisks) into the connective tissue. Most of the connective tissue
cells just beneath the epitheilum are positive (arrows) x610. Intestinal fold at
stage 65. No positive cells are seen in the connective tissue cells x490. Barsz20
um. LE larval epithelium; AE adult epithelium; L lumen (Ishizuya-Oka 1996).

MMP-2 Activation by MT-MMP/T IMP complexes with MT-MMP (panel A),
pro MMP-2 associates with the MT-MMP/I'IMP complex and is activated by
MT-MMP (panel B), The MMP-2/TIMP/MT-MMP complex is active as long
as the trimolecular complex is together (panel C). Once MMP-2 breaks free it is
immediately inactivated by TIMP.

Mmpl4 (A, D, and E) and Timp2 (B) are expressed in the media of muscular
arteries. A-C, show the cardiac outflow tract at 12.5 days p.c. Note hybridization
of the aorta (a) but not of the heart (h) in panels A (Mmp14) and B (Timp2). C,
shows the cardiac outflow region hybridized to the Mmpl4 sense probe. D,
abdominal aorta (a) at 12.5 days p.c. hybridized to Mmp14 antisense probe.
Note labeling in the media (see enlarged box alongside at upper right) and the
absence of labeled cells in the endothelial layer. v, vertebral column; m, posterior

..

Figure 7.4.

Figure 7.5.

Figure 7.6.

Figure 7.7.

abdominal mesenchyme. E, umbilical artery at 17.5 days p.c. hybridized to Mmp14
antisense probe. Boxed area is enlarged alongside (shown at lower left of figure)
showing that labeling is restricted to the media. Arrowheads in panels D and E
indicate the endothelium (Apte 1997).

Hybridization of Mmpl4 probe during urogenital development. A, transverse
section through the developing urinary bladder (b) at 14.5 days p.c. showing
labeled cells in the muscular wall (the lumen of the urogenital sinus is not seen in
this section). Note also labeling of umbilical arteries (a) B, labeled cells are seen
in and around the bulbourethral gland (b) adjacent to the bulbar urethra (u). C
and D, expression of Tiran in ganglia of the 17.5 day mouse embryo. C, dorsal
root ganglia (g); D, trigeminal ganglion (Apte 1997).

MMP-14 and TIMP-2 are expressed in developing synovial joints. Shown are
knee joints of 17.5 days p.c. embryos hybridized to an an14 (A) or Timp2 (B)
antisense probe. Hybridization is seen with cells of the posterior cruciate ligament
(PCL), anterior cruciate ligament (ACL), and patellar tendon (PT). Note labeling
of cells in superficial layers of articular cartilage in both femur (f) and tibia (t)
and the absence of labeled cells in cartilage elsewhere (Apte 1997).

Expression of Mmp14 (A) and Timp2 (B) at the placental implantation site (12.5

days p.c.). Similar regions at the periphery of the implantation site are shown. m,
myometrium; p, placenta (labyrinthine portion). The labeling appears in the
junctional area that includes the decidual reaction of the uterus and
spongiotrophoblast of the placenta (Apte 1997).

Western irnmunoblot analysis of MMP-2 in stromal cell cultures. Samples of 600

pl (A) or 400 pl (B) of 2 d-conditioned media were concentrated, electrophoresed
under nonreducing (A) or reducing (B) conditions, and analyzed by Western
irnmunoblotting with anti-MMP-2. Position of the molecular weight markers
(MW) is indicated on the left margin. (A) Stromal cells (ES) were cultured 14 d
in serum-free medium as described in Fig. 1, in the absence of steroids (C, lane
a), in the presence of 10 nM E +1 uM P (EP, lane b), or received 10 nM E +1
11M P for 14 d followed by steroid withdrawal for 4 d (W, lane c). Human skin
fibroblasts (SF, lane (1), human endometrial epithelial cells (EE, lane e), and
HEC-lA cells (lane f), received no steroids. (B) Stromal cells were cultured 14
d inserum-free medium containing 10 nM E + 1 11M P followed by 6 d in the
presence of 0.2 11M P (P), 10 nM E + 0.2 11M P + 1 11M RU486 (EP + RU486),
or nosteroids (W), and conditioned media collected 2, 4, and 6 d after the time
of withdrawal. (C) Integrative densitometry of the 72-kD MMP-2 band on
Western blots of 2 d—conditioned media collected 2, 4, and 6 d after the time of
withdrawal as described for B. Levels of immunoreactive MMP-2 (IR.MMP-2)
in conditionedmedia were estimated running in parallel known amounts (10-5 00
ng) of purified MMP-2. Values are the meaniSEM (bars) of measurements on
two blots (Irwin 1997).

Figure 7.8. Expression of MMP-2 mRN A in human endometrium throughout the menstrual
cycle. Samples of total RN A(20 pg) extracted from the following tissues were
analyzed by Northern blot hybridization: early, mid, and late proliferative
endometrium (lanes a-d); mid cycle endometrium (MC, lane e); early, mid, and
late secretory endometrium (lanes f-j); early pregnancy decidua (Dec, lane k);
term placenta (Pla, lane 1). (A) Autoradiograph of hybridization with the MMP-2
cDNA probe. (B) Autoradiograph of the same membrane hybridized with the -
actin cDNA probe. (C) Densitometric analysis of the autoradiographs shown in
A and B, using the integrated area under the respective absorbance curves to
calculate the MMP-2/-actin hybridization ratio for each sample (Irwin 1996).

Figure 7.9. Northern blot analysis of MMP and TIMP RNAs during rat skin wound healing.
Total RNA (10 ng) from normal skin (lanes 1) and skin wounds on days 1, 3, 5,
7, 10, and 14 after cutaneous incision (lanes 2-7), were electrophoresed,
transferred to nylon membranes, and hybridized with 32P-1abeled cDNA probes
for rat MTl-MMP, GelA, 8T3, STl, GelB, C013 (A); TIMPl, TIMP2, TIMP3
(B). Blots were reprobed with the 36B4 cDNA used as a loading control (Okada
1997)

Figure 7.10. Quantification of mRN A for cytokine and matrix metalloproteinase in synovia
of mice treated with AP-l DNA oligos or control DNA oligos. mRNA of mice
is reversibly transcribed and coamplified relevant genes are quantified with
reference to coamplified G3 PDH mRN A using Amplisensor assay. The amount
of mRN A is expressed with an arbiu'ary scale for individual cytokines and MMPs
using serial dilution of one representative sample as standard. The differences
between two groups, except for MMP-2 and IL-1 , are < 0.02, and those of IL-
1 are < 0.06 by the Student’s nonpaired T test (Shiozawa 1997).

Figure 7.11. Matrix MMP-2 and MT-MMP expression in Caco-2 cells. (Top) Gelatin
zymography of proteins secreted into the cell culture medium. Lane 3
demonstrates the result following treatment with sulindac sulfide (SS; 25 micro
M) for 24 hr. A 68- and 62-kDa band are seen in the COX-2-expressing Caco-2
cells; whereas, only the inactive 68-kDa band is een in the control Caco-2 cells.
(Middle) Irnmunoblotting results with an anti-MMP-2 antibody in an identical
experiment to that shown in the Top panel. This result confirms the gelatin
zymography data shown in cells which is inhibited in lane 3 by treatment with 25
microM sulindac sulfide for 6 hr. (Tsujii 1997).

4-HPR

AP-l

APMP

bFGF

BPH

Cat

cDNA

CT

cys

DNA

ECM

EGF

LIST OF ABBREVIATIONS

N-(4-hydroxyphenyl)-retinamide
activator protein transcription factor
amino phenyl mercuric acetate

all trans retinoic acid

basic fibroblast grth factor
benign prostatic hyperplasia
collagen-like domain

catalytic domain

complementary DNA

connective tissue

cysteine

deoxyribonucleic acid

extracellular matrix

epidermal growth factor

furin consensus site

fibronectin-like gelatin binding domain
glutamic acid

glomerular mesangial cell

hinge sequence

hemopexin-like domain

his

NGF-B
PA
PAI
PDGF
PEA-3
PIN
PKC
Pre
Pro

pro

TGF-a
TIMP
TNF-ot
TPA
uPA

Xaa

histidine

hydroxyproline

matrix metalloproteinase

nerve grth factor beta
plasminogen activator

plasminogen activator inhibitor
platelet derived grth factor
polyomavirus enhancer A binding protein
prostatic interepithelial neoplasia
protein kinase C

pre-enzyme domain

pro-enzyme domain

proline

retinoic acid

rough endoplasmic reticulium
ribonucleic acid

tumor grth factor alpha

tissue inhibitor of matrix metalloproteinase
tumor necrosis factor alpha

1 2-o-teradecanoylphorbol- l 3 -acetate
urokinase-type plasminogen activator

any amino acid

CHAPTER 1

COLLAGEN CLASSIFICATION

ABSTRACT

Current research on collagens concentrates on the newly discovered types of
collagen. Researchers scramble to uncover their molecular and supramolecular structure,
metabolism, and biological functions. Several pathologies are associated with collagens
in the extracellular matrix. Further elucidation of the structure and functions of the
extracellular matrix will help to provide valuable tools for the management of many
human diseases such as Alport’s Syndrome (hereditary progressive glomerulopathy),
Nail Pattella Syndrome (hereditary disease involving abnormal collagen structure in the
basement membrane), Osteogenesis Imperfecta (brittle bone syndrome), atherosclerosis
(hardening of the arteries), and cancer. This chapter examines collagen through the
discussion of collagen structure, the classification of various types of collagen and the
exploration of the biosynthesis of collagen. Subsequent chapters will focus on the
regulation of collagen degradation in an effort to understand the common mechanisms
involved in tissue remodeling, tissue destruction, and tissue invasion in a variety of normal

and pathological conditions, especially cancer.

INTRODUCTION

Collagen is a protein that is found throughout the animal kingdom. It is the most
abundant protein in humans and other mammals. The term collagen is derived from the
Greek words “kolla” and “genos” which mean “glue” and “formation”. The characteristic
fibers of collagen are responsible for its functional integrity in tissues such as bone, skin
and tendon. As an integral part of the extracellular matrix, collagen provides a structural
framework in tissues. A great many human conditions, both normal and pathologic,

involve the body’s ability to repair and regenerate this collagenous framework.

The first studies in collagen research were performed by chemists, and involved
hide tanning and gelatin production. Collagen was one of the first proteins to be
investigated by x-ray diffraction by Wyckoff and Corey (1936). Modern research on
collagen at the molecular level began in the 1950s. Highberger (1950) and Schmitt
(1953) used the electron microscope to characterize the molecule. Ramachandran and
Kartha (1954) and Rich and Crick (1961) developed models for the triple helical structure
of the protein. Schmitt developed the model of quarter staggered arrays in the native
fibril in 1956. The discovery of procollagen molecular forms, by Bellamy and Bomstein
(1971) initiated the elucidation of the complex multistep biosynthesis of the collagen
molecule. In 1971, Miller and Matukas, through the use of electron microscopy, identified
three distinct but homologous types of collagen. The types of collagen identified were:
type I collagen, the most abundant of collagens and originally thought to be the only
type; type II collagen, shown to be the major protein associated with cartilage; and type
HI collagen, which is found in association with type I collagen in various ratios in most
tissues throughout the body. These types I, II, and III collagens are considered to be the
classical fibril forming collagens. In 1971, Kefalides (1973) discovered a specific type
of collagen associated with basement membranes. This new type IV collagen was
distinguished from the fibril forming classical collagens by its close association with
basement membranes (Timpl 1981). Further identification of various types of collagen
led to the need for better classification based on their structure and function.

This chapter begins its discussion of collagen with a detailed description of the
collagen molecule. Structural information is important for the classification of the various
collagen types. A classification scheme for the various collagen types, which includes
functional roles and distribution, is presented. The complex biosynthesis of collagen is

outlined to establish a general understanding of the genetics and biochemistry of collagens.

Once the structure, classification and biosynthesis of collagen have been described, the
basic knowledge of collagen regulation and degradation of collagens in relation

extracellular matrix turnover will be examined in subsequent chapters.

STRUCTURE OF THE COLLAGEN MOLECULE
The structure of collagen is critical to the function of collagen. The unique

physical characteristics of collagen are conferred by this structure. Classification of the
various collagen types is based on variations in this structure. This section provides a
general overview of the structure of collagen, and discusses how structure is relevant to

the firnction and classification of collagen.

Collagen triple helix

The ability to withstand proteolytic cleavage is one of the most important features
of collagen. This feature is conferred by the collagen triple helix. Figure 1.1 represents
the triple helical structure of the collagen molecule (N imni 1988). All collagen molecules
are characterized by this collagen triple helix, which is composed of three polypeptide
(It-chains. The three (It-chains are coiled in a right hand direction to form a rope like
molecule. The principle feature of the helical domains is the presence of a large number
of glycine (gly) residues. Approximately every third amino acid in an a—chain is a glycine.
This can be followed by any amino acid, but is often followed by proline (pro) and then
by a hydroxyproline (hyp). The repeating sequence of gly-pro-hyp makes up about ten
percent of the type I collagen molecule (Piez 1984). The presence of a glycine every
third residue is absolutely critical to the formation of the triple helix. Glycine is the only
amino acid which lacks a side chain. With the three-chain interaction, each chain alternates
a glycine to form a shallow helix that runs down the middle of a super helical structure
(Kucharz 1992). Side chains of other amino acids are pushed away from the center of

the superhelical structure. The result of this three-chain interaction is the symmetrical

A53— EZV .mEebmmg on 2:: 53 30:6: 22.: 2:. .2662 ofiobw some 2838 mEobmwé 9w 85 em 2329:
05 20% .2335 .3255 ofi 5 2.2.6.2 v25 bo>o @282 ”Earn—w < .2829: cows—Bo 2: mo ogogm Bozo: 2&5 2:. .2 Raw:—

mo_o< 02.}. >4hz<zioowma O
wz_o>1_0 O

 

 

pattern of a left-handed supercoil that possesses a pitch of approximately 8.6 nm (Figure
1.1). The approximate distance between amino acids is 0.291 run, so each glycine residue
is separated by an expansion of 0.873 nm. Individual residues are nearly fully extended
in the collagen molecule (3.6 Angstroms) and this prevents interchain bonds from forming
(Piez 1988). The triple helical structure and amino acid composition makes a collagen

molecule resistant to proteolytic cleavage.

Nomenclature for or—chains

Collagen a-chains can have distinct primary sequences. Figure 1.2 shows the
nomenclature scheme for collagen type 1. In this case the collagen type 1 triple helix is
formed by two identical a-chains and one unique a-chain. The Arabic numeral 1 identifies
the two identical a-chains, designated as al (I) chain (Figure 1.2). The third chain is designated
the a2(I) chain. The Roman numeral in parentheses indicate that these chains are encoded by
genes associated with collagen type 1.

Type I Collagen

/

(011(1))2 (12(1)

alpha chain two copies of alpha chain
primary this chain primary
sequence #1 sequence #2

Figure 1.2. Nomenclature for (IL-Chains of Type I Collagen

Collagen non-helical domains

Although the collagen molecule is characterized by the oc-chain triple helix,
collagen can also contain non-helical domains. Variations in both a-chain primary
structure and domain composition help to establish the various collagen types. Collagens
can have long uninterrupted triple helical domains or the triple helix can be interrupted
by any number of non-helical (globular) domains of various size. Figure 1.3 shows a
schematic representation of different types of collagens and their various helical and
globular domains (N imni 1988). All of the collagen types have a collagen triple helix
associated with their macromolecular structure. The structure of the collagen
macromolecule is related to the function of collagen. In the classic fibril-forming collagens
non-helical domains are positioned on either side of the collagen triple helix structure.
In some types of collagens short non-helical portions interrupt the helical structure.
Therefore, a-chain composition and domain structure used to classify various types of

known collagen.

CLASSIFICATION OF COLLAGENS

Generally collagens are classified into two classes, fibril-forming and nonfibril-
forming collagens. Table1.1 shows the fibril forming collagens and the non-fibril forming
collagens. The fibril forming collagens have continuous or near continuous helical
domains. This feature affects the way they aggregate into fibrils. The non-fibril forming
collagens have helical domains that are separated by variously sized non-helical globular
domains. These features allow these molecules to aggregate into non-fiber forming
fibrous structures.

Currently there are 13 types of known collagen. Each newly purified collagen
type is assigned a Roman numeral to designate its order of discovery. These collagens
Show distinct genetic as well as structural and functional differences. Table 1.2 shows
some common types of collagen, identifies their or-chain composition and their

distribution.

.Awme 88:6 288 3 3258828? 8526 0.8 Eomusnmnmmu
.22: EB 8888 2C. $226 was 25.. 8228 5.38 2: 8 “088 8882 23208 65 mo 62 05 98 $226
2:... 222 82.8 32.88 22:88 38 82:22:88 22: 8 SE 3582 2a 81522: 32388? 32:8 2: mo
.86an .8286 88 28 e83 8888 832w 28 Eve :86 wee 28$ 32 63.5 58:8 868:8 .2 658

OO 59:902er

 

 

 

 

 

T. 5: co. .L
II. 0 . ”£qu6 5326 till-III... 5.2x 2;»
nU lull «Essen .uozoeiuat. ell-IIIIIIII. 5:; .x 2::
call-0'1 :x oa>k

m

4 cl 9' x. oa>h
I. O :> 2;»
F . O .> 2::
0" >_ 01??
.56 .x .> 39:

 

oUo :_ .: ._ 3%:

 

 

Table 1.1. Fibril Forming Collagens and Non-fibril Forming Collagens

 

 

F' '- 11 e N -fi '-F ° Coll
type I collagen type IV collagen
type H collagen type VI collagen
type III collagen type VH collagen
type V collagen type VIII collagen
type XI collagen type X collagen
type XIII collagen
(Associated with fibril forming
collagens)
type IX collagens
type XII collagens
Table 1.2. Chain composition and distribution for some human collagens
wager; CM’n commsition Distribution
Type I (a1(I))2, 012(1) Skin, tendon, bone, cornea, fibrocartilage
Type II (a1(II))3 Dentin, dermis, tendon, hyaline cartilage
Type 111 (0:1 (III))3 Large vessels, uterine wall, dermis, intestine
Type IV (a1(IV))2, «2 (IV) Basement membranes

Each of these collagens shows distinct genetic structural and functional differences.
As stated earlier, the three or-chains in a collagen molecule may have primary sequence
variation among a type or genetic variation between types. These variations are indicated
by Arabic numerals in the case of a primary sequence variation among a collagen type,
and Roman numerals bounded by parenthesis in the case of genetic variation across
collagen types. For example, type I collagen is made up of two a1(l) and one (12(1).
The classification of this family of proteins segregates them into three classes based on
general structural criteria. New collagens are being discovered and their classification is

complex.

Fibril-forming collagens

This class of collagens includes some of the first discovered collagens. Types 1,
II, 111, V, and XI collagens are in this class. These collagens are banded fibril forming
collagens. All members of this class have lengthy, uninterrupted collagenous domains.
As previously described, members of this group are evolutionarily related and show a
common primordial gene unit. These collagens can be separated into two sub-categories,

those that are found in cartilage tissue and those that are non-cartilaginous.

Collagen in cartilaginous tissues

Type II collagen is considered to be the major collagen associated with cartilage
but its distribution is not limited to cartilage. Type II collagen molecules are made up of
three identical alpha chains. These chains are similar to the on] chains of type I collagen,
showing similar electrophoretic characteristics. The cartilage collagen chains are
designated 011(11), where the II is indicative of type II collagen. The most significant
characteristic of cartilage collagen is its high hydroxylysine content and its glycosidically
bound carbohydrates. Collagens from cartilage can contain up to five times more

carbohydrates than the homologous collagen from the skin. This modification is thought

to limit fibril growth of the collagen aggregates (Huerre 1986). Type II collagen is
synthesized during the chondrogenic stages of development from the mesodenn. In the
process of amphibian limb regeneration, the tissues undergo three distinct stages of
development during the transition from mesenchyme to cartilage to bone. A similar
pattern is also seen in chondrogenic activity (Mayne 1990).

Type XI collagen is also found in cartilage and is characterized by three separate
distinct (It-chains that migrate slightly slower than or] (I) chains on SDS-PAGE. Suspected
functions of type XI collagen include regulation of the diameter or growth of type II
collagen fibrils. There is evidence to suggest that the banded cartilage collagen fibers
are polymers of type XI and type II, where type X1 is more concentrated at the fibril
interior than at the surface. This layout indicates that type XI collagen is expressed early

in the formation of cartilage (Mayne 1993).

Collagens in non-catrilaginous tissues

Type I and II collagens are the most common collagens found in non-cartilaginous
tissues. Type I collagen was the first mammalian collagen discovered. It is composed of
three chains, two identical 011(1) and one (12(1). Type I is most abundant in skin, tendon,
ligament, bone, and cornea, where it comprises between 80 and 90% of the total collagen.
Type I collagen is also the primary collagen associated with the bone matrix (N imni
1988). The proportion of collagen in a particular tissue can vary at different sites during
development, aging and pathology. Decreased levels of type I and type IV collagen has
been seen in keratoconus cornea, a condition where the cornea undergoes abnormal
grth and development that leads to impaired vision (Ihalainen 1986). In this disease,
the outer lens of the eye, or cornea is malformed due to the thinning of the corneal
periphery. The resulting cone shaped cornea can cause severe refractive errors. This
disease is most likely the result of decreased collagen content or impaired collagen

maturation.

10

Mutations in the 011(1) and (12(1) genes identify the importance of the normal
type I collagen. The effect of insertions and deletions usually result in phenotypic changes
in bone development characteristics, for example, in osteogenesis imperfecta (01). In
general, the gene defects involving 01 lowers the amount of type I collagen associated
with the bone matrix. This may be the result of intracellular degradation of defective
collagen molecules or a failure of defective molecules to become normally incorporated
into growing fibers (Kucharz 1992).

The molecule that is formed by three 011(1) chains is called the type I trimer
collagen. It is stable and occurs in small quantities in various tissues. The type I trimer
collagen is relatively resistant to fibroblast and neutrophil collagenases. The biological
role of type I trimer has yet to be established. This collagen is synthesized under
pathological conditions. Synthesis of the type I trimer is an early event in the
de-differentiation of chondrocytes in culture. This may occur as a result of the tissue
culture environment. The potential of observing type I trimer synthesis as a measure of
abnormalities in collagen biosynthetic pathway maybe of significant interest (Mayne 1990).

Distribution of Type III collagen is normally associated with the skin where it
accounts for 10-20% of the total collagen. It is also present in varying amounts in
association with type I collagen in lung, heart muscle, heart valves, nerves, blood vessels,
spleen, kidney, lymph nodes, bone and eye. Blood vessels are particularly rich in type
III collagen, where 20-30% of the total collagen in the human aorta seems to be type III.
The amount of collagen type III, in relation to type I, correlates with the extensibility of
the tissue. This generalization has been seen in studies examining changes in the human
aorta during atherosclerosis, where type III collagen shows a marked increase with
atherosclerosis (Murata 1986). Type III collagen is composed of three identical chains.
The pepsin resistant portion of 011(III) is similar in size to the (11(1) and (12(11) chains

and show similar migration on SDS-PAGE (Burgeson 1992). Type III collagen shows a

11

relatively high degree of hydroxylation of proline and a high glycine content (more than
33% of the residues). The most significant characteristic of type H1 collagen is the
presence of intermolecular disulfide bonds involving two cysteine residues close to the
C—terminal region of the triple helix. These intermolecular disulfide bridge formations
allow collagen type III to rapidly cross-link (Burgeson 1992). Such a rapid cross linking
would be of great advantage during early development and wound healing, where collagen
is deposited at a rapid rate.

Ehlers-Danlos type IV syndrome is associated with a gene mutation in the collagen
type III gene. This result in decreased flexibility and increased fragility of the skin and
micro-traumas result in wounds and atrophic scars. Patients with this disorder are subject
to internal organ involvement including mitral valve regurgitation, vessel ruptures,
gastrointestinal abnormalities, and ocular changes. (Kucharz 1992)

Types IX and X11 collagens do not form fibril on their own but they are associated
with fibril—forming collagens. These types do not appear to form supramolecular
aggregates alone, however are often seen to participate in the formation of fibrils in
conjunction with fibril forming collagens.

In cartilage, type IX molecules are localized in a periodic manner along cartilage
collagen fibrils and are in fact cross-linked to collagen type II molecules within such
fibrils. Type IX collagen is assembled by three (1-chains, (11(IX) a2(IX) and a3(IX).
Studies have indicated an inhibitory effect on lateral growth of type H collagen in cartilage,
by type IX collagen (Shimokomaki 1990).

Non-fibril forming collagens

While half of the known collagens forms banded fibers or are associated with
banded fibers, other collagens form structurally and functionally distinct supramolecular
assemblies. The collagens that comprise class 3 have very different structure and function

from the other two classes and from each other.

12

 

 

Type IV collagen is the major component of basement membranes. Basement
Membranes are specialized structures that underlie epithelia and endothelia and perform
many structural and functional roles, which include cell differentiation and orientation,
cell polarization, selective permeability to macromolecules and barriers to cell movement.
Other molecules that form the basement membrane include glycoproteins such as laminin,
entactin/nidogen, proteoglycans, and other less well defined structures. Type IV collagen
has two (11 (IV) and one (12(IV) chains. Recent evidence suggests that type IV collagen
forms homotrimers of either the (11(IV) or the a2(IV) chain (Oikawa 1989).

Type IV collagen does not form well-ordered microfibrils or “Segment Long
Spacing (SLS) Crystals” as do the classic collagens. The characterization of their
supermolecular structure remains elusive, partly due to the amorphous structure of
basement membranes (Glanville 1987). The type IV collagen molecule is characterized
by a long helical domain with several short interrupts. The amino-terminus has a short
helical (78) domain separated from a long triple helical domain (TH) by a small globular
domain, and the carboxy-terminus possesses a globular non-collagenous (NC) domain
known as NC1 (Figure 1.4) (Kucharz 1992). Two models of the supermolecular structure
of type IV have been suggested and are diagrammed in Figure 1.5. These models are
consistent with x-ray crystallography structural observations of type IV collagen (Glanville
1987). Two super molecular structures have been proposed for type IV collagen, the
network model is dependent on the formation of tetramers at the 7S domain, before
dimer formation at the NC1 domain. Subsequent NC1 binding results in a meshwork
formation. The hexagonal model is dependent of the formation of the NC1 dimer before
the 7S trimer, subsequent 78 binding results in the hexagonal meshwork formation
(Kucharz 1992).

Genetic mutations affecting type IV collagen often result in glomerulonephropathy.
For example in Alport’s Syndrome, there is a defect in type IV collagen, the result of

this defect is a systemic thickening of the basement membrane. Most of the research has

13

8883 R23: 282 w82 2:

a

88822..

.303 223st 8880: Cozy genome—39:0: 838% 8882.2 2: :5 at
U 2: 8 Amt 8880: 82—0: 20% 2: 2:32? 8222: 2:an:8 coma—Bo >_ 25—. 26.— oSwE

 

p
4

14

Monomer

/\

Tetromer Dimer

l

 

Network Model Hexagon Model

Figure 1.5. Proposed models for type IV collagen super structure (Glanville 1987).

focused on the basement membrane of the glomerulus. Patients with Alport’s Syndrome
often have thick basement membranes associated with the glomerulus of the nephron.
This thickening is due to a genetic abnormality in type IV collagen, which result in a
malfunction of the filtering activity of the glomerulus. In many cases this will lead to
renal failure in patients with Alport’s Syndrome.

In Nail-patella Syndrome there is an abnormal processing of type IV collagen
that results in systemic pathology. Skeletal and ocular abnormalities as well as renal
involvement characterize this syndrome. Banded collagen fibril formation is seen in the
basement membrane of the glomerulus (Kucharz 1992). Considering its multiple

functions, the malformation of basement membrane has broad reaching implications.

COLLAGEN BIOSYNTHESIS

Collagen Biosynthesis is a complex multistep process. This process was initially
identified with the discovery of the procollagen molecule. The unique structures of
collagens in the extracellular matrix are the result of a number of co-translational and
post-translational modifications of the pre-procollagen molecule. Co-translation refers
to the hydroxylation of an amino acid after a peptide bond is formed but before the
protein is completely formed. Some of these events occur within collagen-producing
cells, and are unique to collagen. Other modifications involved in collagen synthesis are
common and are also seen in other examples of secretory protein biosynthesis. The
following discussion considers the biosynthesis of fibril-forming collagens. Other

genetically distinct collagen types employ slightly different mechanisms of biosynthesis.

The procollagen model
The formation of the intricate network of collagen fibers requires the synthesis
of a precursor. This precursor is known as procollagen and the molecule that immediately

precedes this molecule is known as pre-procollagen (Figure 1.6). Once the procollagen

16

Pre-procollagen

i

Procollagen

l
Collagen

Figure 1.6. The collagen synthesis pathway.

molecule is formed, this molecule can be enzymatically trimmed of its non-helical ends
giving rise to a collagen molecule. The newly processed collagen molecule then
spontaneously assembles into its native fiber in the extracellular space. Procollagen
molecules have been identified for the three interstitial collagens (types I, II, and III)
and many of the amino terminus (N -terminal) and the carboxyl terminus (C-terminal)
peptides (propolypeptides) have been characterized and their primary sequences
determined.

A schematic representation of the type I procollagen molecule is presented in
Figure 1.7. Its synthesis is required for the proper formation of the intricate network of
collagen fibers found in the extracellular matrix (ECM). Procollagen ensures delivery of
the collagen molecule to the site of assembly, by preventing its premature assembly into
fibers in improper locations. The non-helical terminal extensions of type I procollagen
is shown in Figure 1.7. These extensions are later enzymatically cleaved, giving rise to
a collagen molecule that can spontaneously assemble into fibers in the ECM. Type I
collagen has ((11 (1))2 012 (1) chain arrangement. Procollagen is formed by the precursors
pro (11 (I) and pro (12 (1). Both of these propeptide chains have non-helical (globular)
terminal ends. These extensions play a key role in the assembly of the triple helical

procollagen molecule.

17

A33 :80»va 82w 58200
Ease—w 30:02-80: 2: :80 .8080: 30:0: 2: 82.82 :08: 2: 22323 80223003 2 25.3 828800.082 2.08028 < 82 2:me

8&2 388.2- Z
8&2 38822-0 .

INlell-l' A . (0006

 

 

o . ll 0 0102

 

18

The C-terminal ends of procollagen propeptides of both the pro (1-1 and pro 01-2
chains have non-triple helical globular regions associated with them. These globular
peptides have asparagine-linked oligosaccharide units composed of N-acetylglucosamine
and mannose. This carbohydrate side chain is attached near an asparagine-Xaa-threonine
(Asn-Xaa-Thr) sequence (where Xaa is any amino acid). This Asn-Xaa-Thr sequence is
compatible with the structural requirement for glycosylation of asparaginyl residues by
oligosaccharide transferases. The functional importance of this carbohydrate at the
C-terminal is unknown, but it may be a recognition mechanism for alignment, secretion
or assembly into microfibrils (Burgeson 1992). The central segment of this molecule
has an a—helical arrangement and globular regions at both ends. The globular extensions
in procollagen molecules play a key role in the assembly of the triple helical collagen
molecule. Once the procollagen molecule assembly is completed and it is translocated
to the cell surface, the extensions are enzymatically removed from some collagens by
procollagen peptidases and then the fibrillogenesis begins. The procollagen peptidases
are enzymes involved in the selective cleaving of these extensions. Subsequent
spontaneous fibrillogenesis has been observed in many connective tissues and in culture
media derived from collagen-secreting cells (Matuk 1980).

Events leading to the formation of the procollagen molecule and its subsequent
processing can be divided into two categories, intracellular events, and extracellular
events. Intracellular events are those processes that lead to the formation of the
procollagen molecule before its secretion into the extra cellular space. Procollagen
molecular processing continues after its release into the extracellular space and these
modifications are referred to as extracellular events. This next section discusses the
processing of the procollagen molecule and separates them into intercellular events and

extracellular events.

19

Intracellular events
Collagen Genes

Significant progress has been made in the elucidation of the genetic heterogeneity
involved in the formation of the procollagen molecule in recent years. The heterogeneity
and the intron-exon arrangement of the gene family has been found to be more complex
than expected for a molecule as uniform and regular as collagen. Collagen genes are
identified by the letters COL with numbers following separated by the letter “A”, which
refers to the (1 chain. The first number codes for the type of collagen and the second
number codes for the number of the chain. For example, COL3A1 represents the 011
chain of type III collagen.

The human collagen genes that code for the classic collagens have been
characterized through the use of cloned DNA probes. The genes have been mapped to
specific chromosome and these locations are summarized in table 1.3. These studies
localized the gene COLlAl to chromosome 17 (Huerre-Jeanpierre 1986). The other
(1-chain associated with the type I collagen, the COL1A2 gene, was found on chromosome
7 (Henderson 1983). COL2A1 codes for type II collagen the major protein associated
with cartilage and is found on chromosome 12 (Huerre-Jeanpierre 1986). The gene for
Type III collagen resides on chromosome 2 (Kucharz 1992). COL4A1 codes for the
011(1V) chain in type IV collagen, the basement membrane collagen and the gene has
been located on chromosome 13 (Emanuel 1986). The localization of these collagen
genes is useful when considering biosynthesis, expression and regulation of the various
types of collagens.

The size of collagen genes varies from 18,000 base pairs in the COLlAl gene of
humans to 38,000 base pairs of the COL1A2 in the chicken (Kucharz 1992). The gene
structure consists of several exons separated by the non-transcribed introns. The chicken

COL1A2 gene contains at least 52 exons. These exons are 54 base pairs (bp) in length

20

and are separated by large introns that range 80- 2,000 bp. The conservation of base
pairs suggests these exons developed from a primordial gene unit (Burgeson 1992). A
possible model for the assembly of the ancestral collagen gene from a primordial gene
unit in which the 54 bp exon codes for 18 amino acids, the minimum number required for
the formation of a stable triple helical structure, is provided in Figure 1.8 (Kucharz
1992). A schematic representation of the human COLlAl gene is displayed Figure 1.9
(Kuhn 1986). This gene is an example of the conservation of a primordial gene unit.
Twenty-one exons are 54 bp long copies of the primordial gene unit. Nine exons are 108
bp and appear to be a pair of primordial gene units adjoined side by side. Five of the
exons are modified primordial gene units, they show a deletion of 9 bp and have a length
of 45 bp, and the last five exons are duplicates of the modified primordial gene unit and
have 90 bp. The 9 deleted base pairs in the modified primordial gene unit correspond to
one triplicate unit Gly-pro-hyp (Kucharz 1992). The related size of each of these exons
in both the chicken and human suggests that these genes developed from a basic primordial
gene unit (Kucharz 1992).

Regulation of the expression of the collagen gene is controlled at both the
developmental and tissue-specific levels. Regulation of these genes continues to be
elucidated. The promoter for the COLlAl has been characterized (Brenner 1989) and
evidence suggests that angiotensin-converting enzyme (ACE), angiotensin II and
bradykinin are associated with the expression of the type I collagens (Weber 1995).
Further studies in this area will help to provide a greater understanding of normal and

pathological human conditions, such as, development, tissue fibrosis and wound healing.

21

ASSEMBLY OF THE ANCESTRAL COLLAGEN GENE

EXON
I INTRON f 54 Dose poars l INTRON

\ 1 r

T _ . . _ - _ _ . - . . - . - ’7
‘\‘91YL"'Y'9éY ‘ y 9.3, F Y 9&7 ‘ Y 9? ‘ Y 9” ‘1”
I

 

\
I
\
\ ’

‘\ ‘54 b.g.”’
(+) In

AMPLIFICATlON
BY
RECOMBIN AT ION
. WITHIN INTRONS

~-c:1—-o———-—-c:>—cr——o—c> {Jr-“etc.
54 54 54 54 54 54 54

 

Figure 1.8. Diagram of ancestral Collagen Gene and a possible mechanism of collagen
evolution (Kch 1992).

22

S'end S'end triple helix
1 36
222 11031 195 35 102 72 45

   

 

      

 

 

 

51 $0 4.2.81.7 ‘5 6561-43 ‘2 £1 “39 ‘5
38 37
99
21 22 232‘2526 ‘5
2° 111 1 l 99
1, l 111 l l 1
991.5 99 99
“7° 3'end
Hy1930 triple helix 3‘end
16 '
15 152 283191 243114“ 1424
ll.
13 1 .
‘2 11109 8 7 6 s I. 3 2 q 1

 

Figure 1.9. Schematic diagram of the human collagen type 1 alpha 1 chain (Kuhn 1986).

23

Translational, eo-translational, and early post-translational events

The intracellular sequence of events leading to the formation of the procollagen
molecule are shown in Figure 1.10 (N imni 1988). Steps 1-8 show the formation of the
procollagen molecule in the rough endoplasmic reticulum (RER). This begins once a
gene that codes for a specific collagen type is transcribed to yield a functional mRN A. A
specific mRN A is translocated to the RER and is translated by membrane bound polysomes
(Figure 1.10, step 1-3).

Important co-translational modification accompanies translation of the mRNA
(Figure 1.10, steps 3-4). The hydroxylation and enzymes involved in these steps are
shown in table 1.3.

It is a well known observation that neither hydroxyproline nor hydroxylysine can
be incorporated into a peptide before peptide bond formation. Therefore, hydroxylation
of proline and lysine must occur subsequent to peptide bond formation. After the peptide
bond is established, hydroxylation can occur and is mediated by two enzymes, lysyl
hydroxylase and proyl hydroxylase in the presence of iron and ascorbate. The
hydroxylation of proline at both the 3rd and 4th carbon positions is shown in Figure 1.11.
These enzymes are highly specific and act only on chains in the non-helical conformation.
This provides a reason for the time lapse between peptide synthesis and protein folding.

This time varies considerably across various types of collagen. These differences in

Table 1.3. Hydroxylation Enzymes Involved in Collagen Biosynthesis (N irnni 1988)

 

MAC/id Hydroglase Smbol for Figrge 10
proline 3-proline hydroxylase ( . )

proline 4-proline hydroxylase (A)

lysine lysyl hydroxylase (O)

24

  
 
 
  
   

 

_T- 02-mRNA
l .

 

 

 
   
 

     

g ENDOPE PT lDASE

-’ ©

'3

$2

5..

1.11

I:

8

'4

g
.4.

5 ”III? ____

_1 ® ”marina. v.90 ‘45 «aw-aw

O — —

0 ”339.0039.” I’O‘."V‘
.1.

 

T RANSPORT
VESICLE

 

PROCOL LAGEN
C PEPTIDASE

E XTRACELLULAR SPAC E

Figure l . 10. The intracellular sequence of events leading to the formation of the procollagen
molecule (Nimni 1988).

25

~N—CH-C-

Hzc\ C“: ‘l'
CH7

PROLINE

1
”HT?
H,C CH7 +

\/
cu,

PROLINE

 

(5001* -?l-— - - C'IOOH
c-o +01 “2:22 Hg /CH2 + CH: +C02
(EH, CHOH IN:

on, COOH
COOH

lo- HYDROXYPROLINE

 

10°" , MM“ 10°“

(:0 + o; ‘ffl’m": Hzc\ /cuoH+ CH2 + co:
CH3 CH2 CH2

cu, ooou
00011

3 - HYDROXYPROU NE

Figure 1.11. The hydroxylation of proline in collagen protein (Kucharz 1992).

26

time, coupled with the mechanism of hydroxylation and glycosylation, provide insight
into the increased levels of hydroxylation and glycoslyation seen in the basement
membrane collagen. They also may be valuable in the elucidation of collagen pathological
states.

Glycoslyation is shown in Figure 1.10, step 4 and 5. Once the pro Ot-chain of the
procollagen molecule is synthesized, sugar residues are post-translationally added to
hydroxylysine. Galactotransferase and glucosyltransferase mediate this glycosylation.
The function of the glycosylation of hydroxylysine is unclear. Studies suggest that this
modification helps to regulate the fibril diameter. Observations on type IV collagen
glycosylation of hydroxylysine indicate that this modification limits the diameter of the
aggregate (Timpl 1981).

The next post-translational event is the removal of the amino terminal signal
peptide (Figure 1.10, step 6) by endopeptidase. This type of cleavage is typical for
proteins secreted as proenzymes. The function of the signal peptide and the enzyme
responsible for its removal remains unclear. It has been suggested that the role of the
signal peptide is to aid in the transfer of the pre-procollagen polypeptide chain across
the RER membrane. It is generally accepted that a “signal peptidase” cleaves this
extension in a nonspecific manner.

Once the amino-terminal signal peptide is removed, the completed a-chains are
released from the ribosomes, and the Ot-chains begin to interact. Three a—chains associate
through recognition of the carboxy-terminal non-helical domains and begin to form
disulfide cross-links (Mechling 1996) (Figure 1.10, step 7). This initiates the formation
of the triple helical supercoil.

With the linkage of the (1-he1ices through disulfide bonds, the procollagen molecule
can fold and form the triple helical supercoil (Figure 1.10, step 8). Then the procollagen
is ready to be secreted into the extracellular space. Procollagen is translocated to the

Golgi apparatus and packaged into segment long spacing (SLS) crystals (Figure 1.10,

27

step 9) and then placed into secretory vesicles. Fusion of secretory vesicles with the cell
membrane and extrusion of the molecule are the last intracellular stages of collagen
biosynthesis. This is accompanied by the removal of the carboxy-terminal non-helical

extensions and part of the amino-terminal non-helical extensions by specific peptidases

(Nimi 1988).

Extracellular events
Microfibrils form quarter staggered arrays

Microfibrils are formed in the process of extrusion. This is the result of a
cross-linking catalyzed by the enzyme lysyl oxidase. The microfibril is made up of five
strands arranged in a quarter staggered array (Figure 1.12)(Burgeson 1992)(Matuk 1980).
Aggregation

The extracellular aggregation of the collagen molecule and formation of the native
fibril is summarized in Figure 1.12 (Burgeson 1992). Fibrillogenesis is a result of the
quarter staggered arrays undergoing further aggregation. Microfibrils undergo lateral

aggregation followed by end-to-end aggregation.

28

.303 8000305 :3: 0228 2: :0 80:08.8: :5 200208 8022—00 2: .20 80:0w2ww0 222—00880 23. .m _ 2 2:22

.8: 88.00

do

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

lm r d Nul
filre b hi 0 a OHHHIHHGE 0300.06 c0028
N l l I l 4 o
lull K OXIA

 

 

 

 

 

 

 

 

 

6.5.30.6 8 8.66.8

T. 880% T limes.

 

 

 

 

 

 

m _ \l r

00.600000 .223 m t n in:

 

 

 

 

 

 

 

 

29

CONCLUSIONS

It is important to examine collagens in order to understand the role of collagenases
as proteolytic regulators of extracellular matrix turnover. Only with a basic understanding
of the collagen substrate can the activities and the mechanisms of their specific proteolytic
enzymes be truly understood. This chapter provided a basic understanding of collagen
as a molecule that plays an integral role in the extracellular matrix. Tissue remodeling
involves a complex interplay of collagens, matrix metalloproteinases and their inhibitors.
Therefore, in subsequent chapters the matrix metalloproteinases (MMP) and their
inhibitors (TIMP) will be discussed. The common mechanisms involving MMPs and

TIMPs seen in many normal and pathological conditions will also be examined.

30

REFERENCES

Brenner, D. A., Rippe, R. A., and Veloz, L.: Analysis of the collagen alpha 1(1) promoter.
Nucleic Acids Res 17: 6055-6064, 1989.

Burgeson, R. E. and Nimni, M. E.: Collagen types. Molecular structure and tissue distribution.
Clin Orthop 250-272, 1992.

Emanuel, B. S., Sellinger, B. T., Gudas, L. J ., and Myers, J. C.: Localization of the human
procollagen alpha 1(IV) gene to chromosome 13q34 by in situ hybridization. Am J
Hum Genet 38: 38-44, 1986.

Glanville, R. W.: Type IV collagen. in: Structure and function of collagen types. Mayne, R.
and Burgeson, R. E., (eds).Orlando: Academic Press, Inc. 43-79, 1987.

Henderson, A. S., Myers, J. C., and Ramirez, F.: Localization of the human 20) collagen
gene (COL1A2) to chromosome 7q22. Cytogenet Cell Genet 36: 586-587, 1983.

Huerre Jeanpierre, C., Mattei, M. G, Weil, D., Grzeschik, K. H., Chu, M. L., Sangiorgi, F.
0., Sobel, M. E., Ramirez, F., and Junien, C.: Further evidence for the dispersion of
the human fibrillar collagen genes. Am J Hum Genet 38: 26-3 7, 1986.

Ihalainen, A., Salo, T., Forsius, H., and Peltonen, L.: Increase in type I and type IV

collagenolytic activity in primary cultures of keratoconus cornea. Eur J Clin Invest
16: 78-84, 1986.

Kucharz, E. J .: The collagens: biochemistry and pathophysiology. Springer-Verlag, Berlin
1992

Kulm, K.: The collagen family-variations in the molecular and supermolecular structure. in:
Connective tissue: biological and clinical aspects. Kuhn, K. and Krieg, T., (eds).Volume
1. Karger Basel. 29-69, 1986.

Matuk, Y. and McCulloch, C.: The formation of collagen and its relation to ophthalmic
diseases (second of two parts). Can J Ophthalmol 15: 111-116, 1980.

Mayne, R.: Collagen types and chondrogenesis. Ann NY Acad Sci 599: 39-44, 1990.

Mayne, R., Brewton, R. G, Mayne, P. M., and Baker, J. R.: Isolation and characterization of
the chains of type V/type XI collagen present in bovine vitreous. J Biol Chem 268:
9381-9386, 1993.

Mechling, D. E., Gambee, J. E., Morris, N. P., Sakai, L. Y., Keene, D. R., Mayne, R., and
Bachinger, H. E: Type IX collagen NC 1 domain peptides can trimerize in vitro without
forming a triple helix. J Biol Chem 271: 13781-13785, 1996.

31

Murata, K., Motayama, T., and Kotake, C.: Collagen types in various layers of the human
aorta and their changes with the atherosclerotic process. Atherosclerosis 60: 251-262,
1 986.

Nimini M.E.: Collagen volume 2 biochemistry. CRC Press, Bocca Raton, FL 1988.

Oikawa, T., Sayama, K., Matsuda, Y., Fujimoto, Y., Iwaguchi, T., and Matsuzawa, A.:
Characterization of two possible forms of type IV collagen from human kidney cortex.
Biochem Int 19: 615-624, 1989.

Piez, K. A. and Reddi, A. H.: Extracellular matrix biochemistry. Elsevier, New York, NY
1984.

Shirnokomaki, M., Wright, D. W., Irwin, M. H., van der Rest, M., and Mayne, R.: The
structure and macromolecular organization of type IX collagen in cartilage. Ann NY
Acad Sci 580: l-7, 1990.

Timpl, R., Wiedemann, H., van Delden, V., Furthmayr, H., and Kuhn, K.: A network model
for the organization of type IV collagen molecules in basement membranes. Eur J
Biochem 120: 203-211, 1981.

Weber, K. T., Sun, Y., and Katwa, L. C.: Local regulation of extracellular matrix structure.
Herz 20: 81-88, 1995.

32

CHAPTER 2

MATRIX METALLOPROTEINASES (MMPs):
CLASSIFCATION AND BIOCHEMICAL PROPERTIES

ABSTRACT

Matrix metalloproteinases are a family of zinc and calcium-dependent proteolytic
enzymes that have the ability to degrade components of the extracellular matrix (ECM) at
physiological pH. The turnover of ECM contributes to the initial steps of tissue remodeling
during morphogenesis, wound healing, angiogenesis, and tumor invasion. Degradation of
the basement membrane type IV collagen by matrix metalloproteinases (MMPs) occurs in
many normal processes involving tissue remodeling, such as, wound healing and tissue
involution and tissue invasion during angiogenesis and inflammation. In addition, MMP’s
have been associated with several pathological conditions involving tissue destruction as in
rheumatoid arthritis, cancer invasion and metastasis. At present there are 14 known members
of the MMP family. These proteins are loosely classified on the basis of their substrate
Specificity. In this chapter a classification of the members of the matrix degrading
metalloproteinases will be presented along with their substrate specificity, domain structure,

and compositional analysis.

INTRODUCTION
Although several other proteases have the potential to degrade components of the

extracellular matrix, matrix metalloproteinases (MMPs) appear to have the greatest
physiological significance in matrix turnover. Matrix metalloproteinases are a family of
secreted and membrane bound proteins that are capable of degrading components of the
extracellular matrix under normal physiological conditions. Table 2.1 lists the human MMPs
and shows some of their biochemical properties (Murphy 1995). MMPs require zinc and
calcium for their activity. Their major proteolytic activity is directed toward important

components of the extracellular matrix, specifically collagen.

33

Table 2.1. Human matrix metalloproteinases (Murphy 1995)

 

Collagenase

Gelatinase A

Gelatinase B

Stromelysin 1

Stromelysin 2

Matrilysin

Stromelysin 3
Metalloelastase

Membrane-Type

MMP NQ.

MMP-l
MMP-8
MMP-l 3
MMP-2

MMP-9

MMP-3

MMP-10

MMP-7

MMP-11
MMP-12

MMP-14

__9_,_£I_)a.lM 1 Wt- We;

55,000
75,000
65,000
72,000

92,000

57,000

57,000

28,000

5 1,000
54,000

63,000

34

fibrillar collagens types I, II, & III,
types VIII, & X collagens,
proteoglycans, and gelatin

gelatins, Non-helical regions
of fibrillar collagens

specific locus type IV
collagen Types V, VII, XI
collagens, elastin

proteoglycan core protein. Non-
helical regions type IV collagen

procollagens I, H, & III, collagenase,
gelatinase B

strong stromelysin-like activity,
elastin

weak stromelysin-like activity
similar to matrilysin, elastin

progelatinase A

The first animal collagenases were described as enzymes responsible for the dissolution
of the tadpole tail (Gross 1962). Although the existence of bacterial collagenase had long
since been recognized, prior to 1962, there was little or no evidence for an animal collagenase.
However, It was well accepted that the rapid turnover of collagen would be essential for
normal development and tissue remodeling. Gross and Lapiere (Gross 1962) developed an
assay for the detection of collagenolytic activity in living tadpole explants. It was suspected
that during the metamorphosis from a tadpole to a frog, degradation of the ECM, involving
collagenase, would occur particularly in dissolution of the tadpole tail. In these studies fresh
tissue explants from tadpole tails were placed on culture dishes filled with a collagen gel.
The gel was synthesized using a C 14-1abe1ed proline. Figure 2.1 shows the dissolution of the
collagen gel in the area of the tissue explant (Gross 1962). Quantification of free labeled
proline confirmed collagenolytic activity. The protease in these experiments was subsequently
identified as interstitial collagenase and is considered to be the prototypic member of the
family of matrix metalloproteinases. The main criterion, by which a protease was identified
as a “collagenase”, was its ability to cleave a peptide bond in the triple helical region of the
native collagen molecule. The ability of collagen to withstand degradation by proteolytic
cleavage provided a clean definition for a collagenase. This basic definition of collagenases

has been modified since the discovery of additional family members of MMPs.

Figure 2.1 Tadpole tissue explant
digesting collagen. The dark area
around the tissue explant is the area of
collagen lysis (Gross 1962).

 

35

Some collagen types have non-helical interruptions in collagen helical domain and
these non-helical sections are susceptible to proteolytic degradation. Proteins that cleave
these domains may be collagenolytic but would not be considered “collagenases” by the
traditional definition as described above. With the discovery of more genetically distinct
collagens and the importance of collagenolytic enzymes in normal and pathological conditions,
a broader classification scheme must be applied when considering collagenolytic proteins.
The modern classification scheme is based on substrate specificity and domain structure. In
this scheme, the “collagenases” are defined as matrix metalloproteinases, which means that
their catalytic activity is associated with the presence of calcium and zinc ions. Several other
metalloproteinases have been associated with ECM degradation and matrix turnover. Many
of these, although not considered to be “collagenases”, have important collagenolytic fimction.
Therefore, they merit consideration in the examination of ECM turnover.

With the discovery of animal matrix metalloproteinases, studies began to focus on
these proteins and their association with various tissue remodeling events characterized by a
dramatic extracellular matrix turnover. These include conditions such as endometrial cycling
(Irwin 1996, Thompson 1994), mammary involution (Li 1997, Roswit 1988, 1992 ) wound
healing ( Okada 1997 Welgus 1980, 1983,), arthritis and cancer (Nakagawa 1994, Ries
1995)

Matrix metalloproteinases are widely distributed in mammals. They have been
identified in tissues as diverse as skin, liver, uterus, and lung. They are typically present at
basal levels throughout the body. Agents such as interleukin-1 (IL-1), tumor necrosis factor
(TNF) and transforming growth factor-or (TGF-or) induce increased expression of MMPs.
(Birkedal-Hansen 1993). Transforming growth factor [3 (TGF-B) and interleukin-4 (IL-4)
have inhibitory effects on MMP expression (Stetler-Stevenson 1993).

Matrix metalloproteinases are secreted proteins and are usually present in a latent or
inactive form (Willenbrock 1993). Initially, the nature of this latent form instigated controversy.

Some investigators claimed that these enzymes were secreted as an enzyme-inhibitor complex,

36

 

while others suggested that the enzymes were secreted as a proenzyme. While there is
evidence for the existence of MMP-inhibitor complexes (Goldberg 1992), several studies
suggest that MMPs are secreted as proenzymes (Murphy 1995).

Matrix metalloproteinases are produced as proenzymes and require proteolytic
cleavage for activation. Several enzymes such as trypsin, plasmin, plasminogen activator,
cathepsin and kallikrein are capable of this cleavage. Tissue inhibitors of metalloproteinases
(TIMPs) are associated with modulation of MMP activity. TIMPs inhibit MMPs by forming
a 1:1 complex. TIMPs are further discussed in chapter 4. MMPs have the ability to
autoactivate. Autoactivation of MMPs can result in accelerated collagenolysis (Jeffery 1985).
In addition, several MMPs are capable of activating other MMPs via proteolytic cleavage
(Matrisian 1992).

Although many proteases are capable of ECM turnover and basement membrane
degradation, MMPs are believed to be the normal and physiologically relevant mediators of
Matrix degradation. There are several reasons for this argument. First, MMPs are secreted
into the ECM and are, therefore, in the proper location for degradation of ECM. Second,
MMPs are active at neutral pH as would be required under normal physiologic conditions.
Third, there are multiple levels at which MMPs are regulated suggesting a tight control over
their proteolytic activity, as would be required in viva. This chapter begins the examination
of MMPs as modulators of ECM turnover by presenting a classification scheme for the

family of MMPs and identifying some of their biochemical properties.

CLASSIFICATION OF MATRIX METALLOPROTEINASES
Traditionally, MMPs have been loosely classified into three classes according to their

substrate specificity, collagenases, stromelysins, and gelatinases. This classification is limited
because of the overlap in substrate specificity and variations in protein structure, however,
this rough classification scheme is still useful when considering MMPs (Table 2.2). A

combination of protein domain structure and substrate specificity now serves as a more

37

precise classification scheme for the MMP family. Figure 2.2 is a diagram of this new
classification scheme and includes domain structure and major substrates for MMPs (Chambers
1997). This section uses both substrate specificity and domain structure in order to introduce
the known human MMPs.

Collagenase:

This class of MMPs has substrate specificity for fibrillar collagens type I, II, and IH.
There are three members of this class: interstitial collagenase, neutrophil collagenase, and
collagenase 3 (Table 2.2). These MMPs cleave native fibrillar collagens at a single locus that
results it two fragments. The first fi‘agment, named tissue collagenase fragment A (TCA), is
3/4 length of the original molecule, while the second fragment, tissue collagenase fragment B
(T CB), is ’/4 of the original molecule. The approximate site of cleavage is between a Gly-Ile
peptide bond in the a1(I) chain and between a Gly-Leu peptide bond in the (11(H) chain.
The location of these peptide bonds is around 3/4 length from the amino terminus (Highberger
1979, Welgus 1982). These enzymes differ slightly in their substrate preference and origin.
Interstitial collagenase is produced by fibroblasts and macrophages and has preference for
type III collagen, while the neutrophil collagenase shows preference for type I collagen and
has neutrophil derivation (Hasty 1987).

Gelatinases

The gelatinases have specificity for both denatured fibrillar collagen and type IV
collagen. The members of this class include gelatinase A and gelatinase B (Table 2.2).
Gelatinase A (MMP-2) is a 72-kDa protein, and is primarily secreted by human skin fibroblasts.
This enzyme can be activated to catalyze the cleavage of a pepsin-resistant portion of type
IV collagen. Its substrates, in order of preference are gelatin, type IV collagen, type V
collagen, fibronectin, and type VII collagen. It does not cleave interstitial collagens (types I,

II, and III) or laminin (Collier 1988, Seltzer 1976). A second gelatinase has a molecular

38

Table 2.2 Traditional Classification of Matrix Metalloproteinases

 

Collagewes
Interstitial Collagenase

(MMP-1)

Neutrophil Collagenase
(MMP-8)

Collagenase-3
(MMP-3)

Matrix Metalloproteinases

Stromelysins
Stromelysin- l

(MMP-3)

Stromelysin-2
(MMP-1 0)

Matrilysin
(MMP-7)

39

Gelatinases
Gelatinase A

(MMP-2)

Gelatinase B

(MMP-9)

cher MMPs
Metalloelastase(s)
(MMP-1 2)
(MMP-l 8)
(MMP-1 9)

MT-MMPs
(MMP-14)
(MMP-15)
(MMP-16)
(MMP-17)

Stromelysin-3
(MMP-1 l)

weight of 92 kDa and is referred to as gelatinase B (MMP-9). Gelatinase B shares a virtually
identical substrate profile with gelatinase A. The 92-kDa gelatinase B contains an additional
54 amino acid proline rich “collagen-like” domain. The 72-kDa gelatinase A (MMP-2) is
expressed in a wide variety of tissues and is often elevated in malignant tumors (N akagawa
1994, Liotta 1979). Gelatinase B is expressed by a variety of cells such as macrophages,
neutrophils, and epithelial cells (Salo 1983).

Stromelysins

This class, based on substrate specificity, has at least three members: Stromelysin-l
and -2, and Matrilysin. Stromelysin-l (MMP-3) and Stromelysin-2 (MMP-10) are highly
homologous enzymes with molecular weights of 54.0 kDa and 54.1 kDa respectively. Their
major substrates include proteoglycans and glycoproteins, such as, fibronectin and laminin
(Welgus, 1990). Collagen type IV is cleaved by stromelysins at the globular domains only.
A smaller molecule, Matrilysin (MMP-7), is also included in this group because it shares the
same substrate specificity. It has been suggested that matrilysin can cleave native type IV
collagen (Matrisian 1992). Although matrilysin shares a similar substrate profile to the
stromelysins, it is much smaller and has a distinct domain structure, therefore, it may be
useful to apply domain structure when classifying matrilysin (Figure 2.2). The domain structure
of matrilysin is the minimum required for the degradation of ECM by an MMP. This domain
structure includes a leader sequence (PRE), a pro-segment (PRO), and a catalytic domain
(CAT) (Figure 2.2). It has been suggested that this domain structure is most closely related
to the ancestral protein for the family of MMPs (Matrisian 1992).

Other MMPs
Recently, several other MMPs have been discovered that do not fall into the traditional
categories and include, stromelysin-3 , metalloelastase, and Membrane-Type MMPs

(MT-MMPs). Since some of these MMPs have unique substrates or distinct additions to

40

their domain structure, it may be more useful to include domain structure in our classification
scheme. This section identifies some of the new MMPs and briefly discusses their domain
structure in relation to the other MMPs.

Metalloelastase (MMP-12) was first discovered in murine macrophages, and may
represent a new class of elastin degrading MMPs that include MMP-18 and MMP-19. These
MMPs share the same domain structure as the stromelysins and collagenases and have the
PRE and PRO and CAT domains, a hinge segment (H) and hemopexin-like domain (HEM)
(Figure 2.2). Hemopexin is a blood protein that is produced by the liver, its function is to
scavenge free heme in the blood for retrieval back to the liver for processing. Metalloelastase
appears to be more extensively processed at both the carboxy and amino terminal ends.
Active metalloelastase, isolated from macrophages, has a molecular weight of 21 kDa, which
is relatively small compared to the 55 kDa type I collagenase (MMP-1) (Matrisian 1992).

Stromelysin-3 (MMP-11) was first identified in stromal tissue surrounding breast
adenocarcinomas (Basset 1993). Stromelysin-3 will degrade fibronectin and laminin weakly.
The domain structure of stromelysin-3 is distinct from the other stromelysins and collagenases
with the addition of a furin consensus site (F) (Figure 2.2). Furin is a ubiquitous mammalian
Golgi enzyme that is involved in the proteolytic processing of a variety of proteins. The
Furin consensus site has the sequence Arginine -Xaa-Xaa- Arginine (where Xaa is any amino
acid.). This sequence is a recognition site for Furin cleavage. This activation by furin allows
stromelysin-3 to be secreted in an active form.

Membrane-type metalloproteinases (MT-MMP) represent another group of MMPs
that have recently been characterized (Sato 1994). The domain structures of MT-MMPs
also have a furin consensus sites, PRE, PRO, CAT, and HEM domains. In addition, MT-MMPs
have a transmembrane domain (TM) (Figure 2.2). The transmembrane domain results in a
membrane bound MMP. Currently there are four types of MT-MMPs: MMP-l4, MMP-15,

41

Domain Structure for the familiy of Matrix metalloprotienases

 

 

Matrilysin fi Pre I Pro

MMP-7

 

 

Collagenase, E
MMP-- 1 ,
MMP—8, &
MMP- l 3

Pre

Pro 1

 

 

Stromelysins
MMP-3, & E,
MMP-1 0

Pre

 

 

Metalloelastase
MMP-12, E
MMP-18, &
MMP-19

Pre

iPro»

 

 

 

Stromelysin-3 r
MMP-l 1

 

 

MT-MMPs

MMP- l 4, l
MMP- l 5,
MMP-1 6, &
MMP-1 7

Pre

 

 

Gelatinase A E
MMP-2

Pre

 

 

 

Gelatinase B l
MMP-9

Pre

1 as}; 2* 1 EN 11 0111de

 

Figure 2.2. Domain structure for several MMPs. Pre, leader sequence; Pro, prodomain;
Cat, catalytic domain; H, hinge sequence; HEM, hemopexin-like sequence; F, furin recognition
site; TM, transmembrane domain; FN, fibronectin-like binding region; C, collagen-like binding

region.

42

MMP-16, and MMP-17. The major substrates for MT-MMPs have yet to be identified,
however, recent studies have strongly suggested their role in the activation of gelatinase A at

the cell surface (Lim 1996).

BIOCHEMICAL PROPERTIES OF MMPs
All of the MMPs exhibit similar biochemical properties. They are secreted as

proenzymes and require proteolytic cleavage for activation. They are evolutionarily related
and show strong sequence similarity as well as protein domain conservation. All MMPs
require metal ions for their function. This section discusses these biochemical properties of

MMPs and how the differences between the types help to establish their substrate specificity.

Molecular weight characteristics

A comparison of molecular weight estimates in several MMPs is shown in Table 2.1.
These weights are obtained through the use of sodium dodecyl sulfate polyacrylarnide gel
electrophorsis (SDS-PAGE). Stricklin (1977) found MMPs isolated from fibroblasts to have
a molecular weight of 55,000 to 60,000 Daltons (Da) in the proenzyme form and, 45,000 Da
to 50,000 Da in the trypsin-activated form. Similar values have been obtained for rat uterine
collagenase and rabbit synovial collagenase (Roswit 1983). The molecular weight of human
neutrophil collagenase is estimated to be 70,000 to 91,000 Da (Hastey 1986). In most cases
we see a decrease in molecular weight with the activation of MMPs which suggests that the

enzyme is secreted as a zymogen and requires a proteolytic event for activation.

Compositional analysis
Comparison of primary amino acid sequence of various members of the MMP family
demonstrates that these proteins are divided into several groups which possess distinct and

conserved domains among family members. The amino acid sequences have been identified

43

and show a remarkable degree of homology, despite the fact that these enzymes come from
different cell types such as human fibroblasts, human neutrophils, and tadpole tails.

The amino acid sequence of all MMPs, as predicted by the complementary DNA
(cDNA), demonstrates a high degree of conservation between each type of enzyme across
several mammalian species (80% similarity amongst collagenases as well as amongst
stromelysins). There is also a strong degree of conservation between MMP types (50%
sequence similarity) (Murphy 1995). Some of these conserved domains have very specific
functions and are critical for the degradation of ECM. A detailed profile of these domains is

included in the following section.

MMP domain structure

The amino acid sequences of matrix metalloproteinases are well conserved across
several mammalian species (Docherty 1992). The most conserved domains are the PRE
PRO CAT and HEM domains. Figure 2.3 is a schematic representation of the conserved
domains in collagenases, stromelysins and gelatinases (Docherty 1992). The following
discussion describes these domains beginning from the amino terminus and moving toward
the carboxy terminus.

Domain 1 is the propeptide domain and is approximately 80 amino acids in length.
This domain is lost upon activation. As indicated in Figure 2.3, the conserved sequence is
Proline — Arginine — Cysteine - Glycine - (Valine or Asparagine) - Proline — Aspartate
(PRCGV/NPD) (Docherty 1992, Stetler-Stevenson 1989). This domain is responsible for
maintaining enzyme latency. The conserved Cysteine in this region interacts with a zinc
atom bound to the catalytic domain (CAT) or Domain 2.

Domain 2 is the catalytic domain and is just down stream fi'om the pro-domain. This

170 amino acid catalytic domain contains the Histidine - Glutamate - (Phenylalanine or Lysine

44

.aaa €288 £22 a eaoém €88 Beans .6 880% 235:8 .2 2:5

3.503063: S 5.3:...

.33: 20. 2:2. 3 .o 3.3
:0: 532.32: .0 35553-2 2.55.3-2
2:0; 2.56.3-0 azzo>zu< azzsczz
n_nx_ouou
ExgoEoz 5.3 .320Eoz 03:03.5
N 50500

 

 

 

m 50:30 4

f

_c_0Eoo f

 

 

 

IO_\.._\uwI
u u :N 9.552... U
+~

 

 

 

590:3 5.3
32050: 9:32..
m c3800 :225
v 50:30

45

or Isoleucine) - Glycine - Histidine motif (HBF/L/IGH). This motif has significant sequence
similarity to thermolysin, a bacterial metalloproteinase. The histidine residues in this motif
bind the zinc atom, required for MMP fimction. A pair of conserved Asparagine (Asp)
residues located on either side of this domain act as binding sites for calcium, which serves to
stabilize the enzyme.

Domain 3 is located at the carboxy terminus and has sequence similarity to hemopexin,
a heme binding protein, and to the ECM component vitronectin (Matrisian 1992). This
sequence similarity may direct MMPs to their substrates. This sequence is separated from
the catalytic domain by a proline-rich hinge sequence (H) (Figure 2.2). Although this sequence
is not required for catalytic activity, it appears to have a number of important functions. This
sequence is required in interstitial collagenase for the binding and cleavage of fibrillar collagen,
and in gelatinase A, this sequence aids in the activation of gelatinase A at the cell surface.
The two cysteines in this region are thought to form a disulfide bond, resulting in the folding
of this region.

Both gelatinase A and gelatinase B contain a gelatin-binding domain associated with
Domain 2 (CAT), the catalytic domain (Figure 2.3 Domain 4) (Docherty 1992). This domain
has sequence similarity to the gelatin-binding domain of fibronectin. This region is encoded
by the addition of three exons not found in the genes for collagenases and stromelysins. This
region mediates the sequestering of gelatinase within the extracellular matrix.

Domain 5 is found in gelatinase B (MMP-9). This domain is an additional insert
before the hemopexin-like domain and shares homology to the collagen helix (Doherty 1993).
The firnction for this insert is unclear, but it most likely mediates substrate specificity.

As previously stated, matrilysin (MMP-7) has the minimum conserved domains
required for MMP activity. Its major substrates include proteoglycans and glycoproteins,

which are important components of bacterial cell membranes. MMP-7 has a high sequence

46

homology to thermolysin, a bacterial metalloproteinase. It has, therefore, been suggested
that matrilysin most closely resembles the ancestral protein amongst the known MMPs. The
addition of exons to an ancestral gene, which encodes for a matrilysin-like ancestral MMP,
could result in a phenotype with new substrate specificity. It is, therefore, conceivable that

MMPs may have co-evolved with their substrates (Matrisan 1992,Woessner 1996).

Metal ion requirement

The mammalian collagenases are active at optimal pH between 7 to 8. Matrix
metalloproteinases require zinc for their catalytic activity and calcium for stabilization.
Chelating agents such as EDTA and 1,10 phenanthroline effectively abolish enzyme activity.
A reversible loss of MMP activity is observed with the removal of calcium from human skin
fibroblasts, rat skin fibroblasts and rat uterine cells grown in culture. The metalloproteinase
nature of collagenase is evident in several pathological conditions. Ophthalmologists are
able to control the rapid rate of corneal stromal degradation, following alkali burns, with the
use of solutions containing EDTA. Patients with epiderrnolysis bullosa, a collagen disease
that results in loose skin, secrete an abnormal collagenase that has a reduced aflinity for zinc
(Kucharz 1992).

Calcium functions both as an enzyme activator and it also stabilizes the tertiary
structure of the enzyme at physiological temperatures (Seltzer, 1976). Asparagine (Asp)
residues located on either side of the catalytic domain can bind calcium, which stabilizes the
enzyme. Additionally, the inhibition of the MMPs by chelating agents, in the presence of
excess calcium, suggests the participation of a second metal ion in the activation of the
enzyme. Several transition metals have been shown to reverse the inhibition by chelating

agents but zinc was shown to be the most effective (Seltzer, 1976).

47

The role of zinc in metalloenzymes can be divided into four categories: catalytic,
structural, regulatory, and non-catalytic. In the case of matrix metalloproteinases, zinc has a
catalytic firnction. Indeed, zinc is essential for and is directly involved in their enzymatic
activity, and the removal of zinc by chelating agents yields an inactive apoenzyme (Vallee
1992). As previously described, MMPs have a zinc-binding site associated with their catalytic
domains, the I-IEF/L/IGH motif. These binding sites have sequence similarity to the bacterial
metalloproteinase, thermolysin. The zinc atom binds to the Hys residues in this motif. In an
active matrix metalloproteinase this zinc interacts with a water molecule. The water molecule,
upon entering the zinc coordination sphere, is activated for enzyme catalysis by either ionization
or polarization, or is poised for displacement by the substrate, thus making it the decisive
catalytic group (Figure 2.4) (Vallee 1992). In the pro-metalloproteinases the zinc atom
binds with the Cysteine (Cys 73) residue of the pro-domain (Figure 2.5). Only with the

disruption of this bond via the proteolytic cleavage of the PRO domain, or auto perturbation

HO —Boo HOOOH
1 l
Zn Zn
/ l \ H20 / | \
Ionization\ Zln / Polarization
/ | \

S
1
Zn
/ | \
Displacement

Figure 2.4 Schematic function of the H20 ligand in active zinc site of MMP. S, substrate; B,
base (Vallee 1992)

48

 

 

LATENT IS
M42 I / H 000”
9— Zn
\ H
Propeptide
perturbation
b 1
Nth "'"‘
PARTIALLY
ACTIVE
Autocata/ytic
processing
C

NH2

FULLY
ACTIVE

 

 

 

 

Figure 2.5. Schematic diagram of the “Cysteine Switch”, the interaction of the zinc atom in
the catalytic site, with a cysteine in the prodomain of MMP (Docherty 1992).

49

(autoactivation), can the zinc atom be fieed to interact with a water molecule, thereby
activating the enzyme. This mechanism has been described as a “Cysteine switch” and will
be further considered in chapter 3 (Springrnan 1995).

MODULATION OF MMP ACTIVITY
The regulation of collagenolytic activity in the extracellular matrix is clearly important

to organisms. Modulation of IVE/1P activity occurs at several levels including: 1) biosynthesis
and secretion, 2) pro-enzyme activation, and 3) inhibition of MMP by tissue inhibitors of

metalloproteinases. This is the topic of chapter 3.

CONCLUSIONS
The study of collagen degradation, like that of collagen itself, has experienced an

almost exponential growth in the past 20 to 30 years. Initiated by the pioneering work of
Gross (Gross 1962), MMPs are being identified and characterized at a rapid rate. The activity
of MMPs on the extracellular matrix justifies their study in physiological conditions involving
ECM turnover. This chapter has introduced the family of MMPs by presenting a classification
scheme based on protein domain structure as well as substrate specificity. It has also identified
some of the basic biochemical properties associated with MMPs including molecular weight
characteristics, composition analysis, domain structure, and metal ion requirements.
Regulation of MMP activity is of critical importance, since ECM turnover is relevant to
normal conditions such as, development, wound healing and regeneration, as well as
pathological conditions such as arthritis, atherosclerosis, and cancer metastasis. The next

chapter will explore in detail the regulation of MMP activity in these processes.

50

REFERENCES

Basset, P., Wolf, C., and Chambon, P.: Expression of the stromelysin-3 gene in fibroblastic

cells of invasive carcinomas of the breast and other human tissues: a review. Breast
Cancer Res Treat 24: 185-193, 1993.

Birkedal-Hansen, H. J .: Role of matrix metalloproteinases in human periodontal diseases.
Periodontology (5 Suppl) 64: 474-484, 1993.

Chambers, A. F. and Matrisian, L. M.: Changing views of the role of matrix metalloproteinases
in metastasis. J Natl Cancer Inst 89: 1260-1270, 1997.

Collier, I. E., Wilhelm, S. M., Eisen, A. Z., Manner, B. L., Grant, G. A., Seltzer, J. L.,
Kronberger, A., He, C. S., Bauer, E. A., and Goldberg, G. 1.: H-ras
oncogene-transformed human bronchial epithelial cells (TBE-l) secrete a single

metalloprotease capable of degrading basement membrane collagen. J Biol Chem
263: 6579-6587, 1988.

Docherty, A. J., O’Connell, J., Crabbe, T., Angal, S., and Murphy, G.: The matrix
metalloproteinases and their natural inhibitors: prospects for treating degenerative
tissue diseases. Trends Biotechnol 10: 200-207, 1992.

Goldberg, G I., Strongin, A., Collier, I. E., Genrich, L. T., and Marmer, B. L.: Interaction of
92-kDa type IV collagenase with the tissue inhibitor of metalloproteinases prevents
dimerization, complex formation with interstitial collagenase, and activation of the
proenzyme with stromelysin. J Biol Chem 267: 4583-4591, 1992.

Grant, G. M., Cobb, J. K., Castillo, B., and Klebe, R. J.: Regulation of matrix
metalloproteinases following cellular transformation. J Cell Physiol 167: 177-183,
1996.

Gross, J. and Lapiere, C. M.: Collagenolytic activity in amphibian tissue: a new tissue culture
assay. Proc Natl Acad Sci USA 48: 1014-1022, 1962.

Hasty, K. A., Hibbs, M. S., Kang, A. H., Mainardi, C. L., Kawashima, A., Nakanishi, 1.,
Tsuchiya, H., Roessner, A., Obata, K., and Okada, Y.: Secreted forms of human
neutrophil collagenase. Expression of matrix metalloproteinase 9 (92-kDa gelatinase/
type IV collagenase) induced by tumour necrosis factor alpha correlates with
metastatic ability in a human osteosarcoma cell line. J Biol Chem. 261: 5645-5650,
1986.

51

Hasty, K. A., Stricklin, G P., Hibbs, M. S., Mainardi, C. L., and Kang, A. H.: The immunologic
relationship of human neutrophil and skin collagenases. Arthritis Rheum 30: 695-699,
1987.

Highberger, J. H., Corbett, C., Gross, J ., Springrnan, E. B., Angleton, E. L., Birkedal Hansen,
H., and Van Wart, H. B.: Isolation and characterization of a peptide containing the
site of cleavage of the chick skin collagen alpha 1[I] chain by animal collagenases.
Multiple modes of activation of latent human fibroblast collagenase: evidence for the
role of a Cys73 active-site zinc complex in latency and a “cysteine switch” mechanism
for activation. Biochem Biophys Res Commun. 89: 202-208, 1979.

Irwin, J. C., Kirk, D., Gwatkin, R. B., Navre, M., Cannon, P., and Giudice, L. C.: Human
endometrial matrix metalloproteinase-2, a putative menstrual proteinase. Hormonal
regulation in cultured stromal cells and messenger RNA expression during the
menstrual cycle. J Clin Invest 97: 43 8-447, 1996.

Irwin, J. C., Kirk, D., Gwatkin, R. B., Navre, M., Cannon, E, and Giudice, L. C.: Human
endometrial matrix metalloproteinase-2, a putative menstrual proteinase. Hormonal
regulation in cultured stromal cells and messenger RNA expression during the
menstrual cycle [see comments]. J Clin Invest 97: 43 8-447, 1996.

Jeffrey, J. J ., Di Petrillo, T., Counts, D. F., and Cutroneo, K. R.: Collagen accumulation in

the neonatal rat skin: absence of fibrillar collagen degradation during normal
growth. Coll Relat Res 5: 157-165, 1985.

Li, M., Liu, X., Robinson, G, Bar Peled, U., Wagner, K. U., Young, W. S., Hennighausen,
L., and Furth, P. A.: Mammary-derived signals activate programmed cell death

during the first stage of mammary gland involution. Proc Natl Acad Sci USA
94: 3425-3430, 1997.

Lirn, Y. T., Sugiura, Y., Laug, W. E., Sun, B., Garcia, A., and DeClerck, Y. A.: Independent
regulation of matrix metalloproteinases and plasminogen activators in human
fibrosarcoma cells. J Cell Physiol 167: 333-340, 1996.

Liotta, L. A., Abe, S., Robey, P. G, Martin, G R., Roswit, W. T., McCourt, D. W., Partridge,
N. C., and Jeffrey, J. J .: Preferential digestion of basement membrane collagen by an
enzyme derived from a metastatic murine tumor. Purification and sequence analysis
of two rat tissue inhibitors of metalloproteinases. Proc Natl Acad Sci USA. 76:
2268-2272, 1979.

Matrisian, L. M.: The matrix-degrading metalloproteinases. Bioessays 14: 455-463, 1992.

Murphy, G: Matrix metalloproteinases and their inhibitors. Acta Orthop Scand Supp
1266:55-60, 1995.

52

Nakagawa, H. and Yagihashi, 8.: Expression of type IV collagen and its degrading enzymes
in squamous cell carcinoma of lung. Jpn J Cancer Res 85: 934-938, 1994.

Okada, A., Tomasetto, C., Lutz, Y., Bellocq, J. P., Rio, M. C., and Basset, P.: Expression of
matrix metalloproteinases during rat skin wound healing: evidence that membrane

type-1 matrix metalloproteinase is a stromal activator of pro-gelatinase A. J Cell
Biol 137: 67-77, 1997.

Okada, S., Kita, H., George, T. J ., Gleich, G J ., Leiferman, K. M., Ries, C., and Petrides, P.
E.: Migration of eosinophils through basement membrane components in vitro: role
of mauix metalloproteinase-9. Cytokine regulation of matrix metalloproteinase activity
and its regulatory dysfunction in disease. Am J Respir Cell Mol Biol. 17: 519-528,
1997.

Ries, A., Engel, J ., Lustig, A., Kuhn, K.: The firnction of the NC1 domains in type IV collagen.
J Biol Chem 270(40): 23790-23794, 1995

Roswit, W. T., Halme, J ., Jeffrey, J. J ., Seltzer, J. L., Jefl'rey, J. J ., and Eisen, A. Z.: Purification
and properties of rat uterine procollagenase. Evidence for mammalian collagenases
as zinc ion metalloenzymes. Arch Biochem Biophys. 225: 285-295, 1983.

Roswit, W. T., Rifas, L., Gast, M. J., Welgus, H. G, and Jeffrey, J. J.: Purification and
characterization of human myometrial smooth muscle collagenase. Arch Biochem
Biophys 262: 67-75, 1988.

Salo, T., Liotta, L. A., and Tryggvason, K.: Purification and characterization of a murine
basement membrane collagen-degrading enzyme secreted by metastatic tumor cells.
J Biol Chem 258: 3058-3063, 1983.

Sato, H., Takino, T., Okada, Y., Cao, J., Shinagawa, A., Yamamoto, E., and Seiki, M.: A
matrix metalloproteinase expressed on the surface of invasive tumour cells [see
comments]. Nature 370: 61-65, 1994.

Seltzer, J. L., Welgus, H. G, Jeffrey, J. J ., Eisen, A. Z., and Woessner, J. F. Jr: The function
of Ca+ in the action of mammalian collagenases. Matrix metalloproteinases and their
inhibitors in connective tissue remodeling. Arch Biochem Biophys 173: 355-361,
1976.

Springrnan, E. B., Nagase, H., Birkedal Hansen, H., and Van Wart, H. E.: Zinc content and
function in human fibroblast collagenase. Biochemistry 34: 15713-15720, 1995.

Stetler-Stevenson, W. G, Aznavoorian, S., and Liotta, L. A.: Tumor cell interactions with
the extracellular matrix during invasion and metastasis. Annu Rev Cell Biol 9: 541-573,
1993.

53

Stetler-Stevenson, W. G., Krutzsch, H. C., and Liotta, L. A.: Tissue inhibitor of
metalloproteinase (TIMP-2). A new member of the metalloproteinase inhibitor family.
J Biol Chem 264: 17374-17378, 1989.

Stetler Stevenson, W. G, Krutzsch, H. C., Wacher, M. P., Margulies, I. M., and Liotta, L.
A.: The activation of human type IV collagenase proenzyme. Sequence identification

of the major conversion product following organomercurial activation. J Biol Chem
264: 1353-1356, 1989.

Thompson, E. W., Yu, M., Bueno, J ., Jin, L., Maiti, S. N., Palao Marco, F. L., Pulyaeva, H.,
Tamborlane, J. W., Tirgari, R., and Wapnir, 1.: Collagen induced MMP-2 activation
in human breast cancer. Breast Cancer Res Treat 31: 357-3 70, 1994.

Vallee, B.L., and Auld, D.S.: Active binding site of zinc metalloenzyme. Matrix (supp) 1:
5-9, 1992.

Welgus, H. G, Fliszar, C. J., Seltzer, J. L., Schmid, T. M., and Jeffrey, J. J.: Differential
susceptibility of type X collagen to cleavage by two mammalian interstitial collagenases
and 72-kDa type IV collagenase. J Biol Chem 265: 13521-13527, 1990.

Welgus, H. G, Jeffrey, J. J ., Stricklin, G P., and Eisen, A. Z.: The gelatinolytic activity of
human skin fibroblast collagenase. J Biol Chem 257: 11534-11539, 1982.

Welgus, H. G, Jeffrey, J. J ., Stricklin, G P., Roswit, W. T., Eisen, A. Z., Okada, A., Tomasetto,
C., Lutz, Y., Bellocq, J. P., Rio, M. C., and Basset, P.: Characteristics of the action of
human skin fibroblast collagenase on fibrillar collagen. Expression of matrix
metalloproteinases during rat skin wound healing: evidence that membrane type-1

matrix metalloproteinase is a stromal activator of pro-gelatinase A. J Biol Chem
255: 6806-6813, 1980.

Welgus, H. G and Stricklin, G P.: Human skin fibroblast collagenase inhibitor. Comparative

studies in human connective tissues, serum, and amniotic fluid. J Biol Chem 258:
12259-12264, 1983.

Willenbrock, F., Crabbe, T., Slocombe, P. M., Sutton, C. W., Docherty, A. J ., Cockett, M. I.,
O’Shea, M., Brocklehurst, K., Phillips, I. R., Murphy, G, Basset, P., Bellocq, J. P.,
Wolf, C., Stoll, 1., Hutin, P., Limacher, J. M., Podhajcer, O. L., Chenard, M. P., Rio,
M. C., and Chambon, P.: The activity of the tissue inhibitors of metalloproteinases is
regulated by C-terrninal domain interactions: a kinetic analysis of the inhibition of
gelatinase A. A novel metalloproteinase gene specifically expressed in stromal cells
of breast carcinomas. Biochemistry 32: 4330-4337, 1993.

Woessner, J. F. Jr: Regulation of matrilysin in the rat uterus. Biochem Cell Biol 74: 777-784,
1996.

54

CHAPTER 3

MODULATION OF MATRIX METALLOPROTEINASES

ABSTRACT

The turnover of extracellular matrix is regulated by the proteolytic activity
of matrix metalloproteinases (MMP). The regulation of MMPs is modulated at
three levels, expression, activation, and inhibition. The expression of MMP is under
tight regulation and is controlled by a number of growth factors, tumor promoters,
oncogenes and hormones. Proteolytic cleavage is required for the activation of
MPPs and a number of proteolytic enzymes are capable of causing this cleavage,
such as, plasmin, trypsin, and activated MMPs. Tissue inhibitors of matrix
metalloproteinases (TIMPs) can block the degradation of extracellular matrix by
MMPs. By examining the modulation of MMPs, this chapter provides insight into
some of the common mechanisms involved in the turnover of ECM in normal and
pathological conditions. Since MMPs are such a major regulatory force in the

turnover of ECM, a study of their modulation is necessary.

INTRODUCTION

The extracellular matrix (ECM) plays a critical role in normal tissue
development and remodeling. The turnover of ECM occurs during processes
requiring tissue remodeling, such as, embryogenesis, organogenesis, wound healing,
regeneration, and cancer metastasis. The enzymes capable of degrading components
of the extracellular matrix, namely the family of metalloproteinases (MMPs), are
thought to play a regulatory role in these processes. Matrix metalloproteinases are
a family of zinc-dependent endopeptidases which are specific for different
components of the ECM. The major members of this family include collagenases,

stromelysins, and gelatinases. These proteins are secreted as zymogens and, once

55

activated via proteolytic cleavage, are capable of degrading specific components of
the ECM (Howard 1991). Some of the major substrates for MMP are as follows:
collagenases degrade types I, II, and III collagens, gelatinases degrade types IV, V
and VI collagens, and stromelysins degrade laminin, fibronectin, and proteoglycans
(Matrisian 1992). The rate of ECM turnover is determined by the net activity of
the secreted MMP. This activity is controlled at three levels: 1) the rate of production
of the proteolytic enzymes, 2) the activation of zymogens by proteolytic processing,
and 3) the production of specific protease inhibitors (N omura 1989). This multi-level
control meets the requirement of strict regulation for normal function, and implies
that MMP are the physiologically relevant mediator of ECM turnover. The apparent
consequences of abnormal regulation are expressed in several pathological
conditions, such as, rheumatoid arthritis, chronic wounds, osteoporosis, corneal
ulceration, certain genetic diseases and migration of tumor cells through the
basement membrane (Bullen 1995, Schorpp 1995).

Regulation of MMPs is complex and occurs at three different levels;
expression, activation, and inhibition. Several agents including growth factors,
cytokines, oncogenes, UV irradiation and tumor promoters control this regulation

(Angel 1987, Wasylyk 1991).

EXPRESSION OF MMPs

Expression is the first modulation point of MMPs. Increased expression of
MMPs is often seen during tissue remodeling or cellular invasion, as is the case
during development, wound healing, cancer invasion and metastasis. The genes for
MMPs can be induced by a variety of agents including, cytokines, growth factors,
oncogenes, and phorbol esters (Stearns 1993). Individual members of the family of
matrix metalloproteinases do not all respond to the same stimulatory factors. This

point is understandable because not all MMPs need to be expressed at the same

56

time. The expression of any MMP is transient, and is normally under tight regulation.
There is evidence for the conservation of regulatory features in MMP. However,
the transcriptional regulatory mechanisms have not yet been determined for all
MMPs. This section identifies some of the regulatory transcription elements common
to several MMPs, and discusses some of the differences observed in MMP
transcription.

The promoter regions for genes that encode for human stromelysins and
collagenases have important common transcriptional regulatory features (Hagmeyer
1993). Each of these genes has a TATA element associated with its promoter which
is located approximately 30 nucleotides up stream from the transcriptional start
site (Schorpp 1995). Figure 3.1 is a schmatic diagram of the promoter region of an
MMP gene. These genes are characterized as having activator protein transcription
factor binding sites (AP-1) and polyomavirus enhancer A binding Protein 3 sites
(PEA-3). These response sites, AP-l and PEA-3, are the most conserved response
elements in the promoters of MMPs.

The AP-l site is transactivated by AP-l. AP-l consists of a dimerization of
the proto-oncogenes c-fos and c-jun. The AP-l element can be activated by
treatment with the phorbol ester 12-O-tetradecanoy1phorbol-13-acetate (TPA)
(Angel 1987). This induction is due to the activation of protein kinase C (PKC) by
TPA. This activation of PKC results in the activation of c-fos, which results in the
formation of AP-l and the induction of MMP via the AP-l element. Of course
PKC induction does not explain the whole story of AP-l activation. Tumor necrosis
factor a (TNF-or) has been identified as an activator of collagenase through its
prolonged activation of c-jun (Brenner 1989). An analysis of the effects of TNF-a
on collagenase gene expression is presented in Figure 3.2. In these experiments
human fibroblasts, in confluent primary cultures, were treated with TNF-a for 0, 6,

and 24 hours (panel a, lanes 1, 2, & 3 respectively). An increase in the expression

57

680$ :ouatomfib 820:: me $80.“ 20-0 05 3 Baggage 0. 6080—0 88ng m-<mE 0E. 45.50
28 8? mo 5:358? 05 .3 00338353 £ 8080—0 8:302 Tm< 2:. $5820 028%“: m-<mE
98 :2 05 mo 00.00005 2: mega“? .8:% E mo 58an 05 me 883:. 0905058 ._.m BEE

 

 

 

 

 

 

58

 

Figure 3 .2. Human fibroblasts were treated with TNF-a for 0, 6, and 24 hours
(lanes 1, 2, & 3 respectively) in confluent primary cultures. An increase in the
expression of collagenase is seen with prolonged exposure to TNF-or. TNF-oz
activates the AP-l element resulting in an increased expression of collagenase.
Col = collagenase, a—tub = or—tubulin, used as a control (Brenner 1989).

59

   

of collagenase is seen with prolonged exposure to TNF-or. TNF-a activates the
AP-l element resulting in an increased expression of collagenase (Brenner 1989).
Growth factors, such as, epidermal growth factor (EGF) and transforming growth
factor-or (TGF-a) have been shown to elicit proliferative and dedifferentiation effects
by stimulation of proto-oncogenes. Transcription factors c-fos, and c-jun are
activated in the early response to the EGF, and TGF-or cascade of events (Bauknecht
1993). Activation of the AP-l is also mediated by UV irradiation via c-jun. Both
UV irradiation and TPA cause post-translational modification of AP-l within minutes
(Radler-Pohl 1993). The transactivation of the AP-l element alone can lead to
increased expression of MMPs, however, other response elements in the promoters
of MMPs may serve to fine-tune the modulation of MMP expression.

The PEA-3 response element contributes to the fine tuning mechanisms that
modulate the inducibility of the MMP genes by agents such as growth factors,
oncogenes, and cytokines (Matrisian 1992). Another proto-oncogene, c-ets
recognizes the PEA-3 element (Wasylyk 1991). Several growth modulators affect
c-ets activity, including growth factors, non-nuclear oncoproteins, activators of
protein kinase, intracellular calcium and cyclosporin (Wasylyk 1990). Activation
of PEA-3 and AP-l, mediated through c-ets and c-fos/c-jun, results in a synergistic
increase in the expression of collagenase (Wasylyk 1990). In contrast, activation
of the stromelysin promoter by both c-ets and c-jun/c-fos is efficient but the combined
effect is not synergistic. These findings suggest that PEA-3 may serve as a fine
tuning mechanism for differential response to transcriptional regulatory factors.
This is further complicated by the fact that c-ets may serve to activate c-jun and
c-fos (Wasylyk 1991).

It is important to note that gelatinase A lacks the AP-l site which is conserved
in most MMPs. In fact treatment with TPA will actually decrease expression of

gelatinase A. Unlike most other MMPs, gelatinase A is constitutively expressed in

60

_; .r
.: I
' . |

 

a number of cell types. As a result of this basal expression of gelatinase A,
transcriptional elements in the promoter region of gelatinase A have been difficult
to identify. The promoter region of gelatinase A has a noncanonical TATA box site
and two SP1 response elements (Stetler-Stevenson 1996). The noncanonical TATA
sequence in the gelatinase A promoter region has the sequence TACATCT, this
sequence has been suggested to be the basal promoter of gelatinase A (Templeton
1991). In 93% of TATA box consensus sites, position 3 is a T for the canonical
configuration, however there is a 7% possibility for position 3 to be a C, A, or G.
For position 6, A is the base in 83% of TATA-box sequences and there is a 17%
possibility for this position to be held by a C, A, or G. The TACATCT sequence
associated with the gelatinase A promoter differs from the canonical TATA box
sequence at position 3 and 6 only and is considered to be and non-cannonical. Since
the possibility exists for the 6 position to be a C, the TACATCT can be considered
an acceptable, but non-canonical TATA box. In addition, this non-canonical TATA
box site is surrounded by G-C rich sequences, as are most TATA box sequences.
This non-canonical TATA box site allows greater flexibility for polymerase or
necessary transcription factors to recognize the transcription start site (Templeton
1991). Overexpression of gelatinase A may be due to the loss of negative control
circuits. Cell type-specific regulation of gelatinase A has been proposed. The
presence of negative circuits such as DNA methylation or the existence of a silencer
in the DNA has been proposed (Templeton 1991). This cell type-specific regulation
is important since high levels of gelatinase A are associated with pathological
conditions, such as, glomerular nephritis and cancer invasion and metastasis.

In the kidney, gelatinase A is important in inflammation reactions. The over
expression of gelatinase A (MMP-2) in glomerular mesangial cells can lead to an
inflammatory phenotype (Mertens 1997). Recent studies have demonstrated that

glomerular mesangial cells (GMC) exhibit unique stimulation patterns including,

61

 

El

stimulation by interlukin-l, tumor necrosis factor, transforming growth factor-[3
and phorbol esters (Harendza 1995). The transcriptional element associated with
this stimulation is a unique 40 base pair (bp) sequence located at ~1322 to-1282 bp
relative to the translational start site of gelatinase A (Harendza 1995). This sequence
(CTGCTGGGCAAG) was found to interact with the YB-l transcription factor
(Mertens 1997). These findings indicate that YB-l is the cell type-specific regulator
in glomerular mesangial cells (Mertens 1997).

In prostate cancer, gelatinase A is overexpressed in the human prostate
carcinoma cell line, PC-3 cells, after treatment with invasion-stimulatory factor
(ISF). This 78 kDa protein was purified from the conditioned medium of a
bone-metastasizing human prostatic cell line PC-3ML (Stearns 1994). In human
lung carcinoma cells, mRNA for gelatinase A is observed when cells are treated
with the tumor cell derived “collagenase stimulatory factor” (CSF) (Polette 1997).
The mechanism for the transcriptional response to these tumor derived stimulatory
factors has yet to be determined, however, their activity supports the notion of cell

specific regulation of gelatinase A expression.

ACTIVATION OF MMPs
Once the MMP has been secreted into the ECM, the next regulatory stage is

activation of the MMP by proteolytic cleavage. Matrix metalloproteinases are
secreted as latent proenzymes. These latent pro-MMPs require proteolytic cleavage
of the prodomain for activation. The prodomain is a highly conserved region in the
amino terminus, and has the sequence of PRCGVPD. There are several enzymes
that can cause this proteolytic activation, such as trypsin, plasmin, urokinase-type
plasminogen activator, furin (Golgi enzyme), other MMPs, and Cathepsin G. Several

MMPs have the ability to autoactivate and include, collagenases, stromelysins, and

62

 

 

gelatinases. Autoactivation of the 72 kDa progelatinase (gelatinase A) can be
induced by treatment with organomercurial compounds (Howard 1991). A “cystine

switch” is involved in the activation of MMPs.

The cysteine switch

The proteolytic cleavage of the prodomain of the MMP leads to the activation
of the catalytic site of MMP by the “cysteine switch”. The conserved catalytic
domain in MMP binds a zinc metal ion. This zinc ion is the catalytic element of
MMPs. When the MMP is in the proform and the prodomain is still associated with
it, the cysteine in this PRCGVPD sequence interacts with the zinc ion bound to the
catalytic domain. The cysteine and zinc interaction prevents zinc from interacting
with a water molecule, which is required for MMP catalytic function. Proteolytic
cleavage of the prodomain of proMMP leads to the disruption of this cysteine and
zinc interaction, freeing zinc to interact with a water molecule. In the case of
autoactivation there is chemical or mechanical perturbation that causes a disruption
in the conformational structure of the proMMP. This results in the disruption of
the cysteine and zinc interaction. At this point the catalytic domain of MMP is
active and can cleave its own pro-domain as well as that of the surrounding MMPs.
This mechanism, involving the conserved cysteine of the prodomain and the zinc
atom in the catalytic domain described in the above scenarios, has been referred to
as the cysteine switch.

The urokinase-type plasminogen activator (uPA) system is a major contributor
to MMP activation in viva. Plasminogen is a regular component of the plasma and
interstitial fluid and can be converted by uPA to the active enzyme plasmin. Plasmin
is a serine protease capable of cleaving a part of the prodomain of MMPs. The
partial cleavage of the prodomain by plasmin results in the active MMP. Once the

MMP has been activated by plasmin it can cleave the rest of the prodomain by

63

 

autocatalysis. The MMP is then fully activated. Like MMPs, uPA has the AP-l
and PEA-3 elements associated with the promoters of the urokinase-type
plasminogen activator gene (Matrisian 1992). The coordinate transcriptional
activation of collagenase, stromelysin and urokinase may, therefore, result in a
cascade of events that lead to ECM degradation. Basic residues of the “pro”domain
of both collagenase and stromelysin are recognized and cleaved by plasmin. The
subsequent removal of a part of the “pro” domain destabilizes the association of
the cysteine residues in the conserved PRCGVPD region with zinc resulting in a
conformational change. This cascade of events might serve as a control mechanism

to facilitate the total degradation of several components of the ECM.

INHIBITION OF THE ACTIVATED MMP

After MMPs have been expressed and activated, there remains a third level
of modulation. Matrix metalloproteinases can be inhibited by the general serum
proteinase inhibitor a2-macroglobulin (Murphy 1995), however, a family of specific
tissue inhibitors also block the ability of MMPs to degrade the ECM, and appear to
be the most physiologically significant factor. Tissue inhibitors of metalloproteinases
(TIMPs) are a family of ubiquitous proteins and are considered to be the major
players in the regulation of MMP. At present there are at least four known members
of the TIMP family. The classification, structure, and function of these are outlined
in more detail in chapter 4. The primary functional role of TIMPs is to inhibit ECM
degradation by MMPs. This inhibition of MMP by TIMP is mediated by a blockage
of the catalytic domain of MMP. This activity is notably due to the conservation of
12 cysteine residues in TIMP. These cysteine residues form high affinity and

essentially irreversible complexes with the active forms of MMP (Declerck 1991).

64

This interaction of TIMP and active MMP results in the formation of a 1:1 complex
(Declerck 1993). Because of the suspected role in collagen metabolism regulation,
it is important to understand the activity and regulation of TIMPs.

It is interesting to note that TIMP-2 appears to form stable complexes with
procollagenase and progelatinase A. The formation of the procollagenase-TIMP-2
and progelatinase-TIMP-2 complexes prevents the autocatalysis of both pro-forms.
Activation of the procollagenase-TIMP-2 complex can be achieved by proteolysis
of the prodomain by stromelysin. In the case of the progelatinase A-TIMP-2
complex, stromelysin is not an effective activator, however, TIMP-2 appears to
mediate the activation of progelatinase A at the surface of cells by interaction with
membrane type-MMPs (Howard 1991a).

The balance of TIMP to MMP is very important in the turnover of the
extracellular matrix. Any disruption in this balance can lead to pathological
condition, such and arthritis, sclerosis, and cancer metastasis. Like collagen, MMPs
and TIMPs are expressed in a variety of tissues. Transcriptional expression of
TIMP is regulated by a number of factors that, interestingly enough, mirror factors
involved in MMP activation. An AP-l element, responsive to serum and the phorbol
ester TPA, has been identified in the promoter of the TIMP gene. Figure 3.3
highlights the restriction map sequence of the 5’-end of the murine TIMP gene
(Campbell 1991). The TPA-responsive element (TRE) is the element responsible
for the activation by TPA as seen in both the stromelysin and collagenase genes.
This element most likely binds the AP-l protein as observed in MMP activation. In
addition to the AP-I element, a TATA box is also present in the promoter of TIMP.

The TRE element also responds to TGF-B (Campbell 1991). Despite the similarity

65

.A 50 _
fiance—NOV 00mm :80 Reoumtomeeb 05 3 cows—00 E 02098:: 0.8 33:03 =< .08; 0090
05 .3 003000.50.— fl 3?. .0e0w ASE. 05.58 05.? 0:0 .m 05 mo 0oa0=c0m .m.m 0.5me

uuh<00<Uh<hoo<uohbuooheQUbUOhUHtu<o(uuo=uuua<ucb<0huoouoboohh0bohuu
w 0
O

o<3huuotu¢o3hU<<¢<UOOUP5~30UUO<¢OP3<<~O o<uu<u<u<o<o<~uu~uoooou~<
.

L

w
20. 02:3 3%

o<uou~<u~ uuowuohuuuuo<uo»o.—u<uu»ut<.t»<u<8,—(uoUhHUUUOUUhU(<uuouuuu.pt.
u o

0

O
m
IIY TU
<205330Buo<<3<00332532(85850330530£83588qu
‘II o
n
O

OSC-

66

in MMP and TIMP modulation of transcription, MMPs and TIMPs are often
differentially expressed, both spatially and temporally. The fact that TIMPs and
MMP are regulated by similar factors emphasizes a tight level of control over MMP

function.

ROLES FOR MATRIX METALLOPROTEINASES

The complex regulation of MMPs and the importance of ECM turnover
suggest that MMPs play a critical role in many physiological processes. The balance
of MMP to TIMP is important in the spatial and temporal control of ECM turnover.
Several normal and pathological conditions are dependent on this balance.

Extracellular matrix (ECM) remodeling accompanies cell migration, cell-cell
interactions, growth, blastocyst implantation, and tissue invasion during mammalian
embryogenesis. During development, MMPs and TIMPs are differentially expressed
to facilitate cell migration. Mouse embryos secrete functional ECM-degrading
metalloproteinases, including collagenase and stromelysin. These are inhibited as
well as regulated by the tissue inhibitor of metalloproteinases (TIMP) during
peri-implantation development and endoderm differentiation (Brenner 1989). The
expression of collagenase, stromelysin, and TIMP was detected as maternal
transcripts in the unfertilized egg, which were present at the zygote and cleavage
stages, and increased at the blastocyst stage and with endoderm differentiation
(Brenner 1989). These data suggest that metalloproteinases function in cell-ECM
interactions during growth, development, and implantation of mammalian embryos.
MMPs are also associated with several other reproductive processes. Ovulation
and the release of the mature ovum require the activity of MMPs and TIMPs.
Mammary gland involution, following lactation, requires the expression of MMPs

and a decrease in the expression of TIMPs. These findings suggest that MMP and

67

TIMPs are important modulators of ECM turnover required for these processes.
The coordination of MMP and TIMP expression in these conditions is remarkably
similar to wound healing.

In wound healing MMPs and TIMPs are temporally and spatially differentially
expressed. It is likely that the balance between proteolytic enzymes and their
inhibitors, namely neutrophil collagenase and gelatinases A and B, plays a major
role in the maintenance of this critical control. Inflammatory cells secrete neutrophil
collagenase and gelatinase B. Epithelial cells express Gelatinase A. In normal
wound healing, proteolysis occurs in the first few days followed by a shift to matrix
deposition. Studies of chronic wounds show significantly higher levels of gelatinase
A and B enzymes and lower levels of TIMP than those observed in normal healing
wounds (Bullen 1995). Since wound healing requires a balance between synthesis
and degradation of ECM components, it follows that a disruption in the balance
between MMP and TIMP will lead to chronic wounds. Tumors have been described
as wounds that do not heal, and it is expected that these same mechanisms apply to
invasion and metastasis by cancer cells.

During metastasis, there are a series of collagen-containing structural barriers
that a tumor cell must pass. Extracellular matrix and basement membranes must be
breached for cells to intravasate and extravasate. The basement membrane
underlying endothelial cells presents a continuous collagen-containing barrier to
the metastatic process. Within a tissue, at either a primary or a secondary tumor
site, degradation of extracellular matrix must occur in order to permit tumor cell
invasion and spread. Therefore, MMPs are the likely candidates that facilitate the
ECM turnover required by cancer metastasis. Increased expression of MMPs has
been associated with several malignant phenotypes of cancers such as, breast, lung,
skin and the prostate. Several generalizations can be made concerning the expression

of MMPs in these malignant phenotypes: 1) the number of different MMP family

68

members that can be detected tends to increase with cancer progression; 2) the
relative level of any individual MMP family member tends to increase with the stage
of cancer and 3) MMPs can be made by either tumor cells themselves or, quite
commonly by stromal cells, as a host tissue response to the tumor. The expression
patterns for MMPs, therefore, support a role for these MMPs at later stages of

tumor progression.

CONCLUSIONS

The turnover of extracellular matrix is a critical event in many normal and
pathologic conditions. Matrix metalloproteinases are a family of zinc dependent
enzymes capable of degrading the ECM. This chapter reviewed some of the
MMP-mediated mechanisms involved in modulation of ECM turnover. The activity
of MMPs is modulated at three levels, expression, activation and inhibition.
Disruption of MMP modulation at any of these levels can lead to uncontrolled
destruction of ECM observed in many pathologic states including, arthritis, chronic
wounds, and cancer metastasis. Chapter 4 will examine the third level of MMP

modulation involving TIMPs.

69

REFERENCES

Angel, P., Irnagawa, M., Chiu, R., Stein, B., Irnbra, R. J ., Rahmsdorf, H. J ., Jonat, C., Herrlich,
P., Karin, M., Herr, 1., van Dam, H., and Angel, P.: Phorbol ester-inducible genes
contain a common cis element recognized by a TPA-modulated trans-acting factor.
Binding of promoter-associated AP-l is not altered during induction and subsequent
repression of the c-jun promoter by TPA and UV irradiation. Cell 49: 729-739,
1 987.

Bauknecht, T., Angel, P., Kohler, M., Kommoss, F., Birmelin, G, Pfleiderer, A., and Wagner,
E.: Gene structure and expression analysis of the epidermal growth factor receptor,
transforming growth factor-alpha, myc, jun, and metallothionein in human ovarian
carcinomas. Classification of malignant phenotypes. Cancer 71 : 419-429, 1993.

Brenner, D. A., O’Hara, M., Angel, P., Chojkier, M., and Karin, M.: Prolonged activation of
jun and collagenase genes by tumour necrosis factor-alpha. Nature 337: 661-663,
1989.

Bullen, E. C., Longaker, M. T., Updike, D. L., Benton, R., Ladin, D., Hou, Z., and Howard,
E. W.: Tissue inhibitor of metalloproteinases-1 is decreased and activated gelatinases
are increased in chronic wounds. J Invest Dermatol 104: 236-240, 1995.

Campbell, C. E., Flenniken, A. M., Skup, D., and Williams, B. R: Identification of a serum
and phorbol ester-responsive element in the murine tissue inhibitor of metalloproteinase
gene. J Biol Chem 266: 7199-7206, 1991.

DeClerck, Y. A., Yean, T. D., Lee, Y., Tomich, J. M., and Langley, K. E.: Characterization of
the firnctional domain of tissue inhibitor of metalloproteinases-2 (TIMP-2). Biochem
J 289: 65-69, 1993.

DeClerck, Y. A., Yean, T. D., Lu, H. S., Ting, J ., Langley, K. E., Schorpp, M., Mattei, M. G,
Herr, 1., Gack, S., Schaper, J ., and Angel, P.: Inhibition of autoproteolytic activation
of interstitial procollagenase by recombinant metalloproteinase inhibitor MI/TIMP-Z.
Structural organization and chromosomal localization of the mouse collagenase type
I gene. J Biol Chem. 266: 3893-3 899, 1991.

Hagrneyer, B. M., Konig, H., Herr, I., Offiinga, R., Zantema, A., van der Eb, A., Herrlich, P.,
and Angel, P.: Adenovirus EIA negatively and positively modulates transcription of
AP-l dependent genes by dimer-specific regulation of the DNA binding and
transactivation activities of Jun. EMBO J 12: 3559-3572, 1993.

Harendza, S., Pollock, A. S., Mertens, P. R., and Lovett, D. H.: Tissue-specific
enhancer-promoter interactions regulate high level constitutive expression of matrix
metalloproteinase 2 by glomerular mesangial cells. J Biol Chem 270: 18786-18796,
1995.

70

Howard, E. W. and Banda, M. J .: Binding of tissue inhibitor of metalloproteinases 2 to two
distinct sites on human 72-kDa gelatinase. Identification of a stabilization site. J Biol
Chem 266: 17972-17977, 1991.

Howard, E. W., Bullen, E. C., and Banda, M. J .: Preferential inhibition of 72- and 92-kDa
- gelatinases by tissue inhibitor of metalloproteinases-2. J Biol Chem 266: 13070-13075,
1991a.

Howard, E. W., Bullen, E. C., and Banda, M. J .: Regulation of the autoactivation of human
72-kDa progelatinase by tissue inhibitor of metalloproteinases-2. J Biol Chem 266:
13064-13069, 1991.

Matrisian, L. M.: The matrix-degrading metalloproteinases. Bioessays 14: 455-463, 1992.

Mertens, P. R., Harendza, S., Pollock, A. S., and Lovett, D. H.: Glomerular mesangial
cell-specific transactivation of matrix metalloproteinase 2 transcription is mediated
by YB-l. J Biol Chem 272: 22905-22912, 1997.

Murphy, G: Matrix metalloproteinases and their inhibitors. Acta Orthop Scand 66: 55-60,
1995.

Nomura, S., Hogan, B. L., Wills, A. J ., Heath, J. K., Edwards, D. R., Wasylyk, B., Hahn, S.
L., and Giovane, A.: Developmental expression of tissue inhibitor of metalloproteinase
(TIMP) RNA. The Ets family of transcription factors [published erratum appears in
Eur J Biochem 1993 Aug 1;215(3):907]. Development 105: 575-583, 1989.

Polette, M., Gilles, C., Marchand, V., Lorenzato, M., Toole, B., Tournier, J. M., Zucker, S.,
and Birembaut, P.: Tumor collagenase stimulatory factor (TCSF) expression and

localization in human lung and breast cancers. J Histochem Cytochem 45: 703-709,
1997.

Radler-Pohl, A., Sachsenmaier, C., Gebel, S., Auer, H. P., Bruder, J. T., Rapp, U., Angel, P.,
Rahmsdorf, H. J ., and Herrlich, P.: UV-induced activation of AP-l involves obligatory
extranuclear steps including Raf-l kinase. EMBO J 12: 1005-1012, 1993.

Schorpp, M., Mattei, M. G, Herr, 1., Gack, S., Schaper, J., and Angel, P.: Structural
organization and chromosomal localization of the mouse collagenase type I gene.
Biochem J 308: 211-217, 1995.

Steams, M. E. and Wang, M.: Type IV collagenase (M(r) 72,000) expression in human
prostate: benign and malignant tissue. Cancer Res 53: 878-883, 1993.

Steams, M. E. and Wang, M.: Irnmunoassays of the metalloproteinase (MMP-2) and tissue
inhibitor of metalloproteinase (TIMP 1 and 2) levels in noninvasive and metastatic
PC-3 clones: effects of taxol. Oncol Res 6: 195-201, 1994.

71

Stetler-Stevenson, W. G, Harendza, S., Pollock, A. S., Mertens, P. R., and Lovett, D. H.:
Dynamics of matrix turnover during pathologic remodeling of the extracellular matrix
[comment]. Tissue-specific enhancer-promoter interactions regulate high level
constitutive expression of matrix metalloproteinase 2 by glomerular mesangial cells.
Am J Pathol 148: 1345-1350, 1996.

Templeton, N. S. and Stetler Stevenson, W. G: Identification of a basal promoter for the
human Mr 72,000 type IV collagenase gene and enhanced expression in a highly
metastatic cell line. Cancer Res 51: 6190-6193, 1991.

Wasylyk, B., Wasylyk, C., Flores, P., Begue, A., Leprince, D., Stehelin, D., Stetler Stevenson,
W. G, Hewitt, R., and Corcoran, M.: The c-ets proto-oncogenes encode transcription
factors that cooperate with c-Fos and c-Jun for transcriptional activation. Matrix
metalloproteinases and tumor invasion: from correlation and causality to the clinic.
Nature 346: 191-193, 1990.

Wasylyk, C., Gutman, A., Nicholson, R., and Wasylyk, B.: The c-Ets oncoprotein activates

the stromelysin promoter through the same elements as several non-nuclear
oncoproteins. EMBO J 10: 1127-1134, 1991.

72

CHAPTER 4

TISSUE INHIBITORS OF MNIPs (T IMPs):
STRUCTURE, EXPRESSION, GENETIC
REGULATION, AND FUNCTION

ABSTRACT

Tissue inhibitors of metalloproteinases (TIMPs) are secreted proteins that
have inhibitory effects on the extracellular matrix-degrading matrix
metalloproteinases (MMPs). Both TIMPs and MMPs are key determinants of the
establishment, maintenance and turnover of the ECM, and are directly involved in
normal and abnormal human conditions, such as, development, regeneration, and
cancer. Since the ECM is important to cell communication, TIMPs may play an
indirect role in ECM-associated cell signaling. Recent studies have proposed a
secondary effect of TIMPs on cell proliferation, independent of their MMP inhibitory
activities. This chapter examines the most recent studies on TIMPs, including

structure, function, expression, and their regulation.

INTRODUCTION

Tissue inhibitors of metalloproteinases are currently the subject of
considerable interest in many fields of biological research. They play a key role in
the modulation of ECM turnover by virtue of their inhibitory action on matrix
metalloproteinase (MMP) (Matrisian 1992, Murphy 1995). The primary function
of TIMP is to inhibit the activity of MMP. This is preformed by an interaction of 12
conserved cysteine residues of TIMP with the active sites of MMPs. These cysteine
residues form noncovalent bonds forming a MMP:TIMP complex. One MMP and
one TIMP molecule forms this MMP:TIMP complex. The spatial and temporal
balance of MMP to TIMP is essential for normal function. Any prolonged imbalance

of the MMP:TIMP ratio can lead to pathological tissue destruction as seen in cancer

73

and arthritis, or conversely, to an aberrant ECM accumulation as see in fibrotic
states (DeClerck 1992, Edwards 1996, Gohji 1996). A study examining the mean
serum MMP:TIMP ratio in patients with urothelial cancer is presented in Figure
4.1. This study showed that patients with cancer recurrence had significantly higher
mean serum MMP:TIMP ratios than those patients who had no recurrence (Gohji
1996). This study illustrates the importance of a normal balance between MMP
and TIMP. However, the role of TIMP in the modulation of MMP activity is not
completely understood. In some cases TIMPs play a role in the stabilization of
MMPs, and in others, to the activation of MMPs. Recent studies have also indicated
that TIMPs may have additional functions separate from their MMP inhibition,
including the ability to stimulate cell growth. This chapter reviews TIMPs by
introducing some of the members of this family of proteins, followed by an
examination of the primary function of TIMPs. Next, some of the recent studies on
TIMP are discussed including their controversial mitogenic activity and their clinical

application in cancer treatment.

THE TIMP FAMILY

The growing family of mammalian TIMPs is currently composed of three
members. TIMP-1, TIMP-2, and TIMP-3. These TIMPs differ in their structure,
biochemical properties and in vivo expression, suggesting that each form may have
a specific physiological role. The first member of the TIMP family of proteins is
TIMP-1, a glycosylated 28-kDa protein. The second member, TIMP-2, is a smaller
21kDa protein, which has less glycosylation. There is a 40% amino acid sequence
homology between TIMP-1 and TIMP-2 (Hammani 1996, Williamson 1993).
Although TIMP-3 shows sequence homology to the other TIMPs, initial

74

P<0.0l

 

 

 

 

 

 

 

 

1
F P<0.05 -
NS F 1
e
20 I1 1 l
e
.1
1
q o
- e
0. g
15 . : 1
d
q o
.3 1' I
a I ° 0 :
q. ., ' . «11213.43
0. 9 . e
2 ‘0 d :. :O
o '11-
E: q 5:. O” o
2 3 ”-8.48t413 3::
‘ 3:: 177611.55 : e 4-
, E 0'. :
T :3 I.
E cue-"- '
E 1 ” '
o .0. e
m 5 .. e e
: .lL. .
I
I‘ e
d
0
Superficial ' Recurrence ( -) Recurrence (+1
bladder cancer (n=22) (n=3l)
(11:44)

Advanced Urothelical cancer (n=53)

Figure 4.1. The mean serum MMP:TIMP ratio in patients with urothelial cancer. Patients with
recurrence have a significantly higher mean serum MMP:TIMP ratios than those patients who had
no recurrence (Gohji 1996).

75

characterization of this protein reported a closer relationship to the chicken protein
chTIMP-3 (chicken TIMP). Like chIMP-3, TIMP-3 is localized in the ECM, in
contrast to the freely diffusible TIMP-1, and TIMP-2 (Uria 1994). The amino acid
composition and sequence for TIMP-1, TIMP-2, TIMP-3, and chIMP-3 is displayed
in Figure 4.2 (Uria 1994). The conservation of amino acid sequence in these proteins
is important to their common primary function to inhibit MMPs. All of these TIMPs
are able to inhibit the ECM proteolysis by MMP (Murphy 1995, Sato 1992).

All of these TIMPs show the precise spatial conservation of 12 cysteine
residues. These cysteine residues aid in the formation of interchain disulfide bonds.
These bonds are involved in the protein tertiary structure and folding (Figure 4.3).
(Williamson 1993). The disulfide bonds of highly conserved residues are located in
the following sequence: Cys-1-Cys-70, Cys-3-Cys-99, Cys-13-Cys-124, Cys 127-
Cys-174, Cys-132, Cys-137 and Cys-145-Cysl66. The tertiary protein structure
has two domains, made up of six loops (Murphy 1995, Wilison 1996, Williamson
1993). The N-terminal domain (loops 1-3) of TIMP-1 and TIMP-2 is sufficient for
MMP inhibition (Willenbrock 1993). The first 22 amino acids at the N-terminus
are present in all three TIMPs (Uria 1994). However, it is likely that multiple sites
of TIMPs have the ability to interact with MMPs.

DISTINCT SECONDARY SITES OF MMP/TIMP INTERACTION
Although all TIMPs can inhibit all MMPs, TIMP-2 forms specific complexes

with the latent forms of MMP (proMMP) (Goldberg 1992, Imai 1996, Murphy 1995).
In addition to the catalytic domain of MMPs, TIMPs may bind to other distinct
sites on the MMP. The function of these multiple sites of interaction is not yet fully
understood but these interactions may reflect additional functions for TIMPs in

controlling the stability and activation of specific MMPs. Studies have suggested

76

Figure 4.2.Comparison of the structures of the Timp-3 (top row) and Tirnp-l genes
(bottom row). All features of the Timp-3 gene are indicated in bold type, whereas those
of the Tirnp-l gene are italicized. The exons are represented by the amino acid sequences
they encode and are enclosed in boxes. Conserved cysteine residues are in bold type and
indicated by solid circles. Arrowheads indicate that the splice junction splits the codon
and that the split is after the first nucleotide in this codon. The asterisk indicates the only
published splice site for the TIMP2 gene, this splits the relevant codon of TIMP-2 in an
identical fashion. Other splice junctions are between codons. Potential N-linked
glycosylation sites in TIMP-1 are underlined, while a single site in TIMP-3 is overlined.
The sequences of murine TIMP-3 and TIMP-1 were aligned using the DNASTAR
software with gaps introduced for optimal alignment (Apte 1995).

77

 

 

 

 

 

 

I 2 2 a“
m m m m
C. DT KN.
SS QQ VI
LL PP. LR
LL HH KP
VL SP KS
VL PP fiGG
LI SA VM
GG eCC VF
L. .18 1 m. LKK
w. .CC m m AA
PS AA ~.m RR
TA EK 11
.L AS'VV...
.8 GS IL
.A w. DD
.F H. SS
.P GA NN
.A LI eCC
.M SL FF
MM w. AA

 

 

 

KK,

.T
MM
QM
KK
TK
YY
VR

.Q

.Y
LL
TT
G.
F.

D. .

 

 

 

 

3M
m...
RK EB '66
YL SM TT
MM Av LI
EP LL
TT YF
HY QE
IA YE
YY KS
n QR NR w
13 v1 VN &
Wm H. 3 m
m PA LS m
MA KK
KN LH
86 .A
.V .Y
.A GG
.K eCC
FF LL
06 SS

 

 

 

....intron 4.

exon 4
exon 5

 

HT
DR
ww
RP.
EV
VL
FF.
NS
eCC
LA
GN
Tl.
YH
MF
KK
GG
EN
YR
VL
RR

 

.V
NG
oCC
60
LA

RK
YK
NS
LF.
GV
KR
RQ
QQ
SA
LP
TS
LL

 

......intron 4.....-
intron 5

 

eCC

KT.

 

 

exon 5
exon 6

 

 

ww
LL
eCC
EH
NT
KD
SS
TE
VL
FK
eCC
PP
LI
YS
YL
oCC
SP
KF
[V

0CC
AA
YF
HH
KR
SS
QQ
YY
GD
PE
YS
60
F.
NV
SL
LV
MQ
DD

 

TT.

 

GWAPPDKSI

R

 

 

PA
DG

AS
NR

S .

 

 

78

 

.GVU_"©
m n

9.0L
@6996

ooooooooo
. m

   

m
ooooooooooueooo
o o

VU m _

o. o
61 6
Q m
o. e
mu 1 my
9 1 .
m P ‘e
o w ”W
6 L .m
0 I
©a 0
0 '0
o- m
0 9
© 6
9 w 9.9

w m .®
090000®0®0w®960

Figure 4.3. Protein folding in TIMP-1, showing the conserved cysteines and

formation of loops in TIMP (Caterina 1997).

79

that TIMP-2 has the ability to form complexes with the pro form of type I collagenase
and gelatinase A. In the case of type I collagenase, TIMP appears to stabilize the
enzyme, however, this stabilization is not observed in gelatinase A

The interaction of TIMP-2 and the proform of type I collagenase (proMMP-l)
results in the stabilization of the enzyme. TIMP-2 binds to the hemopexin domain
of proMMP-1. This interaction blocks autoactivation of MMP-2, however, the
proMMP-1 :TIMP-Z complex can still be activated by cleavage by stromelysin. Once
the MMP-lzTIMP-Z complex is activated, a second TIMP (TIMP-1, -2, or —3) is
required for inactivation.

The carboxy-terminus of TIMP-2 will interact with the carboxy-terminus of
gelatinase-A (MMP-2) (Willenbrock 1993). The proMMP-2:TIMP complex is a
product of the C-terminus domain of TIMPs, as show via studies using truncated
proteins that contain only the N-terminal domain (Murphy 1991). This interaction
was shown to yield an increased rate of MMP-2 inactivation by a factor of 100. A
schematic representation of the interaction of TIMP-2 and gelatinase A by their
C-termini is presented in Figure 4.4. This interaction is involved in the activation
of gelatinase A at cell surfaces, it is mediated by MT-MMPs, and is discussed in

greater detail in chapter 6.

TIMP-3 IS BOUND TO THE ECM
A striking feature of TIMP-3 and chIMP-3 is that they are ECM-bound,

whereas both TIMP-1 and TIMP-2 are freely diffusible, extracellular proteins
(Hammani 1996, Uria 1994, Williamson 1993). The restricted localization of these
proteins suggests that they may have the greatest potential to modulate pericellular
proteolysis (Uria 1994). The molecular basis for the ECM association of TIMP-3
continues to be elusive but one interesting possibility has surfaced. Both TIMP-3

and chIMP-3 have motifs composed of seven amino acids flanked by two basic

8O

.222 Lo 3% €38

. 2: 88m 853% 8?. a .3 < omega—om Ea NAHSE. mo 330885 05 mo eouSEmoaoc 238038 < iv 85mm..—

    

EoEwomoE

 

32825 < enacts—ow Ao>u85v < among—om

Eofiwomocm

 

 

 

 

< 82:20on

A

”Am—2:.

 

33826 < amaze—ow

81

amino acids, or B(Xaa7)B motifs, B is a basic amino acid, here lysine (K) or arginine
(R), and Xaa is any other amino acid) These motifs have been implicated as having
a close association with hyaluronic acid (Yang 1994). TIMP-3 has three of these
B(Xaa7)B motifs, one is in the N-terminus domain and two are juxtaposed in the C-
terminus domain. Since these B(Xaa7)B motifs are not found in TIMP-1 and TIMP-
2, a strong possibility exist that TIMP-3 association with the ECM is meditated by
the interactions of this B(Xaa7)B motif and hyaluronic acid or other related

glycosaminoglycans (Yang 1994).

PREFERENTIAL INHIBITION OF MMP BY TIMP
Despite the fact that all MMP family members, by definition, are inhibited by

TIMPs, whether different TIMP species show preferential inhibition of particular
active MMPs, is still very much open. TIMP-2 has been shown to be more effective
at inhibiting both gelatinase-A and gelatinase-B, than TIMP-1 (Howard 1991).
TIMP-1, on the other hand, is more effective in the inhibition of interstitial
collagenase than TIMP-2. TIMP-2 will bind to gelatinases approximately 10 times
faster that TIMP-l (Howard 1991). The preferred inhibition of gelatinase B by
TIMP-3 has also been observed in mouse embryos in the process of implantation.
Studies have identified gelatinase-B as an enzyme that is essential for the cellular
invasion involved in embryo implantation. Gelatinase B is expressed principally by
trophoblast giant cells in implanting (7.5 days post coitum) mouse embryos (Harvey
1995). At this time, there is a high level of expression of TIMP-3 that is restricted
to maternal decidual cells immediately adjacent to the implanting embryo (Harvey
1995). This suggests that TIMP-3 forms a protective shield that is deployed by the
uterus to contain the highly invasive embryo. Thus, TIMP-3 may be the inhibitor
of choice in tissue remodeling situations that involve transient, high level production

of gelatinase-B, the added advantage being that it provides a mechanism for spatial

82

restriction of MMP activity by virtue of its ECM-association. Preferential inhibition
0f the various TIMPs for different MMPs provides a fine tuning mechanism by

Which to control specific ECM turnover.

DIFFERENTIAL EXPRESSION OF TIMP

Additional information regarding the differential expression of TIMPs during
development has provided further evidence for the specific role of these inhibitors.
In adult mice TIMP-1 is preferentially expressed in epithelial tissue, cartilage and
muscle. In contrast, during murine embryonic development, TIMP-3 is preferentially
expressed in these same tissues (epithelial, cartilage, and muscle), while TIMP-1 is

expressed specifically in the bone (Hammani 1996). The phorbol ester,
1 2-O-tetradecanoy1phorbol- l 3-acetate (TPA), can induce the expression of TIMP-1
and TIMP-3. A similar induction of TIMP-1 and TIMP-3 is observed by treatment
with transforming growth factor B (TGF-B). However, TPA and TGF-B down
regulate TIMP-2 production (Khokha 1992, Lim 1996, Wilson 1996). In
macrophages, lipopolysacharides have been identified as up-regulators of TIMP-2
and the down-regulators of TIMP-1 (Lim 1996). Together, these observations

suggest that each member of the TIMP family has a specific physiological role.

REGULATION OF THE TIMP GENES

The in vivo patterns of expression of the three TIMP genes are distinct
(Nomura 1989). In the mouse, the expression of TIMP-1 has been localized to
areas of active ECM remodeling, such as, sites of osteogenesis in the developing
embryo, and the uterus and ovaries of adult females. TIMP-2 shows a distinct
accumulation in the placenta just prior to birth. TIMP-3 mRNA transcripts are
found at high levels in certain adult tissues that lack expression of other TIMPs,

iIlCluding the kidney and the brain (Leco 1994). In the developing mouse embryo,

83

TIMP-3 transcripts are abundant in the surface epithelia of organs, such as, the
developing bronchial tree, kidney, colon and esophagus that have extensive tubular
Structures (Apte 1994). These studies further support the notion that the individual
T IMPs perform specific and non-interchangeable tasks. The patterns of
Stimulus-responsiveness and tissue-specific expression of the three TIMPs differ
dramatically. The transcription sites AP-l, and PEA-3 are found within the promoters
of the many TIMP gene promoters. Interestingly, AP-l and PEA-3 sites are also
found in the promoter regions of several MMP genes. The production of TIMP-1
and TIMP-3 is induced by the activation of AP-l and PEA-3 sites in response to
treatment with TPA or TGF-B (Lim 1996, Fabunmi 1996, Grant 1996). This suggests
that AP-l is a critical factor that may serve to coordinate the responses of MMP
and TIMP genes. In contrast, TIMP-2 is not induced by TPA and TGF-B and has a
very diffrent regulation mechanism (Campbell 1991), indicating the possibility that
different elements are at work in the promoter of TIMP-2. The anti-inflammatory
glucocorticoid dexamethasone strongly induces TIMP-3, whereas it inhibits TIMP-1
and TIMP-2 expression (Leco 1994). A comparison between the 5’-flanking regions
of the three TIMP genes is presented in Figure 4.5. The following section describes
some of regulatory elements associated with the promoters of TIMP genes.

The expression of TIMP-1 is largely confined to the adult bone and the ovary
and to tissues undergoing remodeling or inflammation. In culture, serum and growth
factors modulate the expression of TIMP. The promoter of the TIMP gene contains
approximately six exons and does not possess a TATA box sequence. Elements
associated with the TIMP-1 gene confer a transcriptional activation by viruses,
Serum, phorbol esters, and TGF-B. TIMP-l has two AP-l sites and PEA-3 response
element. The AP-l sites are inducible by interaction with both heterodimers of

cthS/c-jun or homodimers of c-jun. The PEA-3 response element is activated by

c‘ets, however, this element is dependent on AP-l activation. Therefore, this PEA-3

84

.Gofi 888805 080w TEE. 05.88 05 5 885 085 8 8:86 8885 8888 080w TEAS. 8.08::
05 .8 5:0-.m 05 8 w5ecoam0bo0 0080.800 536 < 508055 880 0.2 8855 05 .8 85588 05 55? 85 0.003 a mm 550% 5
88 8580 05 5.5.5 5 $85 .3 508000.808 0.8 m0c0w F322. 0588 88 NiEEb 05 8 8x0 58¢ 05 5:0 080m 732:. 0588 05 5
898 025 8a 05,—. .m080w 085 05 5 050 :80 8855888 85.88 05 .8 85588 05 895 505880505 08 08080—0 085 .8 8558a
05. .080w TASS. 05.88 05 88 080m 7% 05.88 05 5 885 085 55.5 50.88800 .80 080m nigh: 05 .8 80880.8 05 5
8850 08 88080.0 beam—8&0.— 88588898 0588A .0080w m2; 005 05 .8 8880.— mfixceafi 05 8002505 888800 .3. 088,»

85A. “2. 8.8.. _._m_ 3.83 88$

3». 3h. ems. :3. uvnn. ems.

 

 

 

 

 

 

85

response element may have the same fine tuning function as PEA-3 in the collagenase
gene. The interaction of AP-l and PEA-3 elements in the transcription of TIMP-1
suggests that multiple signal transduction pathways coordinate and regulate TIMP-1
production.

The promoter region of TIMP-2 is associated with one AP-l , two AP-2 and
three PEA-3 binding sites. The AP-l biding site appears non-responsive to treatment
with TPA. The position of this AP-l site is relatively close to the transcriptional
start site, and may be non-responsive due to its position. The production of TIMP-2
has been shown to be down regulated by lipiopolysaccharides in macrophages and
upregulated by cAMP in human fibroscarcoma (HT1080) cells. The AP-2 binding
sites have been shown to mediate the cAMP response in many genes and two of
these sequences are found in the promoter of TIMP-2. Lipopolysaccharides have
been shown to affect gene regulation via interleukin 6 (IL-6). IL-6 activates a
nuclear factor (NF) known as NF-IL6 and there are two NF-IL-6 recognition
sequences associated with the promoter of TIMP-2 (Hammani 1996). These
differences of TIMP-2 gene expression are likely to result in the specific role TIMP-2
plays in the inhibition of MMP.

The promoter region of TIMP-3 gene is very extensive. Computer analysis
of cloned TIMP-3 promoter has revealed six AP-l binding sites, two nuclear factor
NF-KB sites, one c-myc site, and two copies of a p53-binding motif separated by
eight base pairs with two mismatches at the second motif, along with many other
cis elements (Sun 1995). TIMP-3 gene expression is inducible by TPA, and TNF-B
just like TIMP-l. These stimulatory agents transcactivate via AP-l and NF- KB
sites. It is, however, possible that the NF —KB site provides the mechanism for the

distinct temporal pattern of expression seen between TIMP-1 and TIMP-3.

86

TIMP EFFECTS ON CELL GROWTH

In addition to its structural role, the ECM is an information-processing
medium that controls cell position, identity, proliferation and fate. This functional
role can be the result of a direct interaction with cells since cells are often dependent
on their contact with the ECM. In addition, the ECM also provides a contextual
framework for the interaction of growth factors and cytokines with their signaling
receptors (Murphy 1995). Given the importance of TIMPs as regulators of the rate
of ECM turnover, it is reasonable to propose that they exert some influence on
ECM-dependent cell signaling.

In the mammary gland, cells depend on instructive signals from the underlying
basement membrane. Are the activities of mammary cells directed by the expression
of TIMP or MMP? In the case of mammary gland, TIMPs are expressed at high
levels during lactation where they promote cell growth and milk production (Talhouk
1992). Increased expression of MMP-2 is seen during mammary involution with a
decrease in TIMP production. By protecting the basement membrane, TIMPs
maintain the ECM integrity required during lactation. Lactation ceases when the
protective role of TIMP is overcome by excess MMPs leading to basement membrane
destruction, and mammary gland involution. Several proteins expressed during
mammary gland lactation and involution such as casein, gelatinase A, gelatinase B,
stromelysin-l , and TIMP. Casein is a milk protein and its expression is indicative
of lactation. The expression of casein (lactation) appears to be protected by the
expression of TIMP against the expression MMPs, initially. However, as the levels
of gelatinase A, gelatinase B, and stromelysin-1 increase the level of casein decreases.
The levels of TIMPs appear to go up initially in response to MMP expression but in
a short time the TIMPs inhibition is overcome, lactation ceases and mammary

involution begins.

87

In all of the situations discussed above, any effects of TIMPs on cell growth
would be exerted indirectly via the ability of TIMP to modulate MMP. Perhaps the
strangest aspect of the family of TIMP is that they may be directly involved in the
control of cell proliferation. This controversial function was first identified in
TIMP—l. TIMP-1 is an Erythroid Potentiating Activator (EPA) and can stimulate
both in vivo and in in vitro growth of erythroid precursors (Hayakawa 1992). Several
subsequent studies have suggested TIMP-1 is mitogenic for a number of cell types
including fibroblasts, keratinocytes and a number of hematopoetic cell types. The
EPA function has also been observed for TIMP-2. A broad-spectrum mitogenic
activity, optimally stimulating the growth of human gingival fibroblasts and Raji
Burkitt’s lymphoma cells at small concentrations, has been proposed as additional
functions of TIMP-2 (Hayakawa 1994). Alklinization of TIMP—1 and —2 blocks
their inhibitory effects on MMPs, but does not affect their growth promoting
activities. This suggests that this activity is distinct from their MMP inhibitory
activity. Progelatinase-A/TIMP-2 complexes, which retain anti-MMP activity
against endogenous active MMPs, are ineffective growth promoters. Since TIMP-
2 interacts with progelatinase A by the carboxy terminus, this finding suggests that
the growth promoting feature of TIMP-2 is associated with the carboxy terminus
(Hayakawa 1994). These studies argue that at least some of the effects of TIMPs
on cell growth involve functions that are distinct from their MMP inhibiting activity.

There is also evidence to suggest that under certain conditions TIMP-l may
inhibit cell growth. Originally, Khokha (1992) suggested the importance of the
maintenance of TIMP balance by showing that an artificially-mediated reduction of

TIMP-1 in Swiss 3T3 cells conferred an oncogenic phenotype (Khokha 1992). Later

88

studies revealed that TIMP-l overexpression in B16FIO mouse melanoma cells
reduced their metastatic abilities, but that this effect might not be solely attributed
to diminished cellular invasiveness (Khokha 1994). Rather, there is evidence for a
suppressive effect of TIMP-1 on the growth of melanoma cells after extravasation
from the circulation (Koop 1994). A study using intravital videomicroscopy, to
measure the time course of extravasation, was used to examine B16Fl 0 cells
transfected with the TIMP-1 gene and the wild type B16F10 cells. It was expected
that the wild type B16F10 cells would extravasate faster because they expressed
MMPs. In contrast, this study showed that both cells extravasated at similar rates.
However, they did see an increased time for the initial division for the B16F10 cells
that were tranfected with the TIMP-1 gene (Koop 1994). This observation indicates
TIMP-l has some inhibitory effects on cell growth.

Nothing is currently known about the effects of mammalian TIMP-3 on cell
growth but chicken TIMP-3 promotes the growth of non-transformed chicken
fibroblasts in low serum conditions. Currently, the mechanisms for both the induction
of cell growth as well as the modulation of this activity are unknown. Elevation of
cAMP levels following activation of adenylate cyclase by a G-protein-coupled
mechanism is suggested to be a potential signaling pathway involved in the mitogenic
effects of TIMP-2 on human HS68 fibroblasts and HT108O fibrosarcoma cells
(Tanaka 1995). Despite these intriguing findings there is still widespread skepticism
concerning the growth promoting effects of TIMPs. This issue is unlikely to be
resolved until TIMP receptors and signal transducers are identified and characterized

at the molecular level.

89

APPLICATIONS

Matrix metalloproteinases play an important role in a number of normal and
pathologic human conditions. TIMPs appear to block the destruction of ECM by
MMPs. TIMPs have been show to block invasion and metastasis (Declerck 1992,
Duncan 1996, Ito 1996, Motgomery 1994, Nuovo 1995). They also have anti-
angiogenic and tumor suppressive properties (Dunsmore 1996). Clinicians and
researches endeavor to apply the knowlege of TIMP to human health care. MMPs
have become targets for anticancer, antiarthritic, and opthalmological therapies.
Several agents, namely tetracyclines and synthetic inhibitors of MMP, such as
batimastat (BB-94), are currently under clinical investigation for the treatment of
periodontal disease, arthritis and cancer.

Batimastat, ({4-N-hydroxyamino}-2R-isobutyl-3S-{thienylthiomethyl
Succinyl}-L-pheny1alanine-N-methylamide) (BB-94) is a member of the rapidly
expanding class of synthetic inhibitors of MMPs. These drugs are designed to block
MMP activity, and are currently under clinical investigation as anticancer agents.
Through its ability to block stromelysin, gelatinase-A, gelatinase-B, and matrilysin,
BB-94 has been shown effective in the inhibition of human cancers. The effect of
BB-94 on breast cancer regrowth, using MDA-MB-443S breast carcinomas in nude
mice, is presented Figure 4.6 (Sledge 1995). The data show that BB-94 has a
suppressive effect on tumor growth. Other studies indicate that treatment with
BB-94 causes inhibition of ovarian cancer cell growth, prevention of Bl6-BL6
murine melanoma and human colon cancer cell metastasis, and inhibition of

angiogenesis (Sledge 1995).

90

Tumor Volume (mmB)

5000 A 1 - A A A A A A L 4 A L

 

 

 

 

 

 

45001 q,- 1.
4000< b
1 0 88-94 1
35001 D Control . r
1 r
3000‘ in p
1 l
J

2500+ 1
20001 - r
I "' _

l
1500‘ b
1 . . 1
10001 0 r
1 - . P
5001 lo .
I i P

0 - . - e , - , - r - j - a
10 11 12 13 14 15 16 17
VVeek

Figure 4.6. Effect of batimastat (BB-94) treatment on MDA-MB-435 primary
tumor regrowth in athymic nude mice. Tumor growth is measured by tumor
volume, treatment with batismastat decreased tumor growth (Sledge 1995).

91

CONCLUSIONS
TIMPs are multifunctional proteins. It is well established that TIMPs modulate

the activity of MMPs by the formation of a one to one complex that interferes with the
catalytic activity of MMPs. The role of TIMP in the regulation ECM turnover is
well accepted and researchers and clinicians are rushing to apply the current
knowledge to modern medicine. Conversely, In some cases, TIMP-2 actually aids
in the activation of progelatinase A. In addition, it has been suggested that TIMPs
play an important role in cell proliferation. It is clear that the maintenance of the
ECM could help to maintain some cell growth or at least inhibit cell death as is seen
in mammary gland involution. However, one controversial observation suggests
that TIMP may have mitogenic function independent of their MMP regulatory
function. Further study of this multifunctional protein will provide important
information that can be applied towards the treatment of connective tissue diseases
as well as cancer. The next chapter examines the roles of MMP and TIMP in cancer

invasion and metastasis.

92

REFERENCES

Apte, S.S., Olsen, B.R., Murphy, G.: The gene structure of tissue inhibitor of
metalloproteinases (TIMP)-3 and its inhibitory activities define the distinct TIMP
gene family. J Biol Chem 270:14313-14318, 1995.

Campbell, C. E., Flenniken, A. M., Skup, D., and “Williams, B. R.: Identification of a serum
and phorbol ester-responsive element in the murine tissue inhibitor of metalloproteinase
gene. J Biol Chem 266: 7199-7206, 1991.

Caterina, N.C., Windsor, L.J., Yermovsky, A.E., Bodden, M.K., Taylor, K.B.,
Birkedal-Hansen, H., Engler, J .A.: Replacement of conserved cysteines in human
tissue inhibitor of metalloproteinases—1. J Biol Chem 272:32141-32149, 1997.

DeClerck, Y. A., Perez, N., Shimada, H., Boone, T. C., Langley, K. E., and Taylor, S. M.:
Inhibition of invasion and metastasis in cells transfected with an inhibitor of
metalloproteinases. Cancer-Res 52: 701-8, 1992.

Duncan, W. C., McNeilly, A. S., and Illingworth, P. J.: Expression of tissue inhibitor of
metalloproteinases-1 in the human corpus luteum after luteal rescue. J Endocrinol
148: 59-67, 1996.

Dunsmore, S. E., Rubin, J. S., Kovacs, S. O., Chedid, M., Parks, W. C., and Welgus, H. G.:
Mechanisms of hepatocyte grth factor stimulation of keratinocyte metalloproteinase
production. J Biol Chem 271: 24576-24582, 1996.

Edwards, D. R., Beaudry, P. P., Laing, T. D., Kowal, V., Leco, K. J ., Leco, P. A., and Lim,
M. S.: The roles of tissue inhibitors of metalloproteinases in tissue remodelling and
cell growth. Int J Obes Relat Metab Disord 20 Suppl 3: S9-15, 1996.

Gohji, K., Fuj imoto, N., Fujii, A., Komiyama, T., Okawa, J ., and Nakajima, M.: Prognostic
significance of circulating matrix metalloproteinase-2 to tissue inhibitor of

metalloproteinases-2 ratio in recurrence of urothelial cancer after complete resection.
Cancer Res 56: 3196-3198, 1996.

Goldberg, G. I., Strongin, A., Collier, I. E., Genrich, L. T., and Marmer, B. L.: Interaction of
92-kDa type IV collagenase with the tissue inhibitor of metalloproteinases prevents
dimerization, complex formation with interstitial collagenase, and activation of the
proenzyme with stromelysin. J Biol Chem 267: 4583-4591, 1992.

Harvey, M.B., Leco, K.J., Arcellana-Panlilio, M.Y., Zhang, X., Edwards, D.R., Schultz,
G.A.: Proteinase expression in early mouse embryos is regulated by leukaemia
inhibitory factor and epidermal growth factor. Development. 121: 1005-1014, 1995.

93

Hammani, K., Blakis, A., Morsette, D., Bowcock, A. M., Schmutte, C., Henriet, P., and
DeClerck, Y. A.: Structure and characterization of the human tissue inhibitor of
metalloproteinases-2 gene. J Biol Chem 271: 25498-25505, 1996.

Hayakawa, T., Yamashita, K., Tanzawa, K., Uchijima, E., and Iwata, K.: Growth-promoting
activity of tissue inhibitor of metalloproteinases-l (TIMP-l) for a wide range of
cells. A possible new growth factor in serum. FEBS Lett 298: 29-32, 1992.

Hayakawa, T., Yamashita, K., Ohuchi, E., and Shinagawa, A.: Cell growth-promoting activity
of tissue inhibitor of metalloproteinases-2 (TIMP-2). J Cell Sci 107: 2373-2379,
1994.

Howard, E. W., Bullen, E. C., and Banda, M. J .: Preferential inhibition of 72- and 92-kDa
gelatinases by tissue inhibitor of metalloproteinases-2. J Biol Chem 266: 13070-13075,
1991.

Imai, K., Ohuchi, E., Aoki, T., Nomura, H., Fujii, Y., Sato, H., Seiki, M., and Okada, Y.:
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,
1996.

Khokha, R., Zimmer, M. J ., Graham, C. H., Lala, P. K., and Waterhouse, P.: Suppression of
invasion by inducible expression of tissue inhibitor of metalloproteinase-1 (TIMP- 1)
in B16-F10 melanoma cells. J Natl Cancer Inst 84: 1017-1022, 1992.

Koop, S., Khokha, R., Schmidt, E. E., MacDonald, I. C., Morris, V. L., Chambers, A. F., and
Groom, A. C.: Overexpression of metalloproteinase inhibitor in B16F10 cells does
not affect extravasation but reduces tumor growth. Cancer Res 54: 4791 -4797, 1994.

Leco, K.J., Khokha, R., Pavloff, N., Hawkes, S.P., Edwards, D. R.: Tissue inhibitor of
metalloproteinases-3 (TIMP-3) is an extracellular matrix-associated protein with a

distinctive pattern of expression in mouse cells and tissues. J Biol Chem. 269:
9352-9360, 1994.

Lim, Y. T., Sugiura, Y., Laug, W. E., Sun, 8., Garcia, A., and DeClerck, Y. A.: Independent
regulation of matrix metalloproteinases and plasminogen activators in human
fibrosarcoma cells. J Cell Physiol 167: 333-340, 1996.

Matrisian, L. M.: The matrix-degrading metalloproteinases. Bioessays 14: 455-463, 1992.

Montgomery, A. M., Mueller, B. M., Reisfeld, R. A., Taylor, S. M., and DeClerck, Y. A.:
Effect of tissue inhibitor of the matrix metalloproteinases-2 expression on the growth

and spontaneous metastasis of a human melanoma cell line. Cancer Res 54: 5467-5473,
1 994.

Murphy, (1, Docherty, A. J ., Hembry, R. M., and Reynolds, J. J .: Metalloproteinases and
tissue damage. Br J Rheumatol 30 Suppl 1: 25-31, 1991.

94

Murphy, G: Matrix metalloproteinases and their inhibitors. Acta Orthop Scand Suppl 266:
55-60, 1995

Nomura, S., Hogan, B. L., Wills, A. J ., Heath, J. K., and Edwards, D. R.: Developmental
expression of tissue inhibitor of metalloproteinase (TIMP) RNA. Development 105:
575-583, 1989.

Nuovo, G J ., MacConnell, P. B., Simsir, A., Valea, F., and French, D. L.: Correlation of the
in situ detection of polymerase chain reaction-amplified metalloproteinase

complementary DNAs and their inhibitors with prognosis in cervical carcinoma.
Cancer Res 55: 267-275, 1995.

Sato, H., Kida, Y., Mai, M., Endo, Y., Sasaki, T., Tanaka, J., and Seiki, M.: Expression of
genes encoding type IV collagen-degrading metalloproteinases and tissue inhibitors
of metalloproteinases in various human tumor cells. Oncogene 7: 77-83, 1992

Sledge, G. W. Jr, Qulali, M., Goulet, R., Bone, E. A., and Fife, R.: Effect of matrix
metalloproteinase inhibitor batimastat on breast cancer regrowth and metastasis in
athymic mice. J Natl Cancer Inst 87: 1546-1550, 1995.

Sun Y, Hegamyer G, Kim H, Sithanandam K, Li H, Watts R, Colbum NH: Molecular cloning
of mouse tissue inhibitor of metalloproteinases-3 and its promoter. Specific lack of

expression in neoplastic JB6 cells may reflect altered gene methylation. J Biol Chem
270: 19312-19319,1995.

Talhouk, R. S., Bissell, M. J ., and Werb, Z.: Coordinated expression of extracellular
matrix-degrading proteinases and their inhibitors regulates mammary epithelial function
during involution. J Cell Biol 118: 1271-1282, 1992.

Tanaka, K., Iwamoto, Y., Ito, Y., lshibashi, T., Nakabeppu, Y., Sekiguchi, M., and Sugioka,
Y.: Cyclic AMP-regulated synthesis of the tissue inhibitors of metalloproteinases

suppresses the invasive potential of the human fibrosarcoma cell line HT1080. Cancer
Res 55: 2927-2935, 1995.

Uria, J. A., Ferrando, A. A., Velasco, G, Freije, J. M., and Lopez Otin, C.: Structure and
expression in breast tumors of human TIMP-3, a new member of the metalloproteinase
inhibitor family. Cancer Res 54: 2091-2094, 1994.

Willenbrock, F., Crabbe, T., Slocombe, P. M., Sutton, C. W., Docherty, A. J ., Cockett, M. 1.,
O’Shea, M., Brocklehurst, K., Phillips, I. R., and Murphy, G: 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,
1993.

Williamson, R. A., Smith, B. J., Angal, S., Murphy, G, and Freedman, R. 3.: Structural
analysis of tissue inhibitor of metalloproteinases-1 (TIMP-1) by tryptic peptide
mapping. Biochim Biophys Acta 1164: 8-16, 1993.

95

Wilson, C. L. and Matrisian, L. M.: Matrilysin: an epithelial matrix metalloproteinase with
potentially novel functions. Int J Biochem Cell Biol 28: 123-136, 1996.

Yang, B., Yang, B. L., Savani, R C., and Turley, E. A.: Identification of a common hyaluronan
binding motif in the hyaluronan binding proteins RHAMM, CD44 and link protein.
EMBO J 13: 286-296, 1994.

96

CHAPTER 5

THE ROLE OF MATRIX METALLOPROTEINASES IN
INVASION BY CANCER CELLS

ABSTRACT

Remodeling of the extracellular matrix (ECM) is a prerequisite to cellular
invasion. Matrix metalloproteinases(MMPs) have been associated with the process
of ECM turnover, and are thought to play a critical role in tumor cell invasion and
metastasis. The overexpression has been correlated with tumor progression in many
cancers. Specifically, the uncontrolled expression of gelatinase A, gelatinase B,
and stromelysin-1 is believed to be critical because of their ability to degrade type
IV collagen, a major component of basement membranes. This review will focus
on gelatinase A and gelatinase B, and the stromelysins (stromelysin-1, -2, -3 and
matrilysin), because of their ability to degrade type IV collagen. This review also
examines MT-MMPs, an integral membrane MMP, that has been implicated in the
activation of gelatinase A. By examining the distribution, expression, and activation

of these MMPs, this review sheds light on the their role in invasion and metastasis.

INTRODUCTION

A benign tumor grows within the confines of a basement membrane, can
progress to a malignant tumor, which traverses the basement membrane, and invades
the surrounding stromal tissue. Further progression of the carcinoma from invasion
to metastasis requires the intravasation into, and the extravasation out of the blood
circulatory and lymphatic systems. These systems serve as highways for the
dissemination of malignant tumor cells to potential secondary, metastatic tumor
sites. The cancer cells must now invade a new site, again crossing basement

membranes to seed themselves into the secondary tissue. In this metastatic site the

97

tumor must grow and establish itself away from its original cellular environment.
The ability of cancer cells to penetrate the barriers of multiple basement membranes
and the stromal extracellular matrix strongly indicates the presence of proteolytic
enzymes capable of extensive degradation of matrix proteins.

The matrix-degrading metalloproteinases (MMPs) are a family of secreted,
zinc and calcium-dependent, proteolytic enzymes that have the ability to degrade
components of the extracellular matrix at physiological pH. MMP activity is thought
to be important in processes that involve tissue remodeling and tissue invasion
(Edwards 1996). In addition, MMPs have been associated with several pathological
conditions involving tissue destruction, including rheumatoid arthritis and cancer
invasion and metastasis (Mignatti 1996). Based loosely on their substrate specificity
the family of MMPs has been traditionally divided into at least three major classes,
which include the collagenases, the gelatinases and the stromelysins (Pajouh 1991).
All of the members of the MMP family possess conserved functional and structural
domains, which translate into similar mechanisms of latency, activation and
proteolysis (Murphy 1995). The amino terminal or activation locus and the catalytic
site that binds the zinc atom are the sites of greatest homology within this family
(Matrisian 1990). Except for matrilysin, the carboxy terminal region of all MMPs
has a hemopexin/vitronectin-like domain that has been shown to be important in
the binding of substrates and inhibitors. A review of the MMP family is presented
in Table 5.1 (Corcoran 1996). The gelatinases have a fibronectin type gelatin-binding
domain which contributes to their substrate specificity. Both gelatinase A and
gelatinase B degrade type IV collagen (Conway 1996). Stromelysins are made up
of four MMPs, three of which are normally secreted by stromal cells (Stromelysins
-l, -2, -3). All of the stromelysins, including matrilysin, have a broad substrate
specificity that includes the non-helical region of type IV collagen as well as

additional components of the extracellular matrix, such as, fibronectin and vitronectin

98

Table 5.1. MMP Family, including subgrouping, nomenclature, mRNA & protein

 

 

 

 

 

 

size and major substrates
Subgroup MMP Nomenclature mRNA Protein Substrate(s)
kb ltD
Interstitial MMP-l Fibroblast 2.0 52.57 Fibrillar collagen (Ill > > I)
collagenase MM P-8 PMN 3.3 75 Fibrillar collagen (l > > >III)
MMP-l3 Collagenase-3 2. 2.5. 3 S4 Fibrillar collagen
Stromelysin MMP-3 Stromelysin-.1 1.9 52. S8 Lam. FN core protein.
nonhelical collagen
MMP-l0 Stromelysin-2 1.7 58 Same as above
MMP-ll Stromelysin-3 2.2 29 a1 Antitrypsin
MMP-7 Matrilysin 1.1 28 Similar to MMP-3
Gelatinases MMP-2 Gelatinase A 3.1 72 Gelatin. Col I and IV, FN, Vn.
elastin. B—emploid
MMP-9 Gelatinase B 2.3 92 Gelatin. Col IV, Col V
MTMMP MMP-14 MTMMP—l 4.5 63 Pro-gel A
MMP-15 MTMMP-Z 3.6 72 ?
MMP-l6 MTMM P-3 2.1 64 Pro-gel A
Elastase MMP-12 Metalloelastase 1.8 53 Elastin

 

(Corcoran 1996).

99

(Conway 1996). Matrilysin, a special member of the stromelysin group, has strong
proteolytic activity and is produced by glandular epithelia (Wilson 1996).
Stromelysins are also important because they have the ability to activate other MMPs
including gelatinase B. Membrane-type matrix metalloproteinases (MT-MMPs)
contain a transmembrane domain which facilitates cell-surface activation of
gelatinase A (Imai 1996). This review will focus on the gelatinases, and stromelysins
because of their ability to degrade type IV collagen and other components of the
basement membrane. Although MT-MMPs have not been identified as MMPs that
degrade type IV collagen, they will be considered in this review because of their

role in the activation of gelatinase A (MMP-2).

DISTRIBUTION OF MMPS INVOLVED IN BASEMENT MEMBRANE
INVASION

The normal distribution patterns of MMP expression divides the family of
MMPs into three distinct categories (Table 5.1). First, there are MMPs that are
primarily expressed by the stromal cells, these include, stromelysins-l, -2, -3, the
interstitial collagenases, and gelatinase A (Crawford 1995, Sato 1992). Second,
inflammatory cells express MMPs including neutrophil collagenase, gelatinase B,
and macrophage metalloelasetase. Expression of gelatinases A and B has also been
observed in epithelial cell types. Third, matrilysin is expressed in normal epithelial
tissue. Matrilysin is the only known MMP constitutively expressed by normal
epithelial cells (Wilson 1996). This section examines the distribution patterns and

substrate preferences of gelatinases, stromelysins and MT-MMPs.

Gelatinase A
A 72 kDa type IV collagenase, called gelatinase A, or MMP-2 (Table 5.1) is

constitutively secreted by human skin fibroblasts and a number of human normal

and tumorgenic cell types. This enzyme catalyzes the cleavage of the a-helical

100

domain of type IV collagen, a major component of the basement membrane. The
substrates for gelatinase A in the order of preference are gelatin, type IV collagen,
type V collagen, fibronectin, type VII collagen, and elastin (Murphy 1993).
Gelatinase A has been isolated from normal skin explants and several tumor cell

types (Docherty 1992).

Gelatinase B
Gelatinase B (MMP-9) is a 92 kDa gelatinolytic enzyme (Table 5.1).

Neutrophils and macrophages typically express this enzyme. Its normal expression
is often associated with the inflammatory phase of wound healing. Overexpression
of gelatinase B has been observed in epithelial cells and several invasive tumor cell
types (Webber 1995). The substrates, in order of preference for gelatinase B are

gelatin, type IV collagen, type V collagen, and elastin.

Stromelysins

Stromelysin-1 and —2 expression has been normally observed during tissue
remodeling. In human wound healing both stromelysins and interstitial collagenases
are expressed at sites of wound healing (Crawford 1996). Stromelysin-1 has been
proposed to be involved in the remodeling of dermis as in the removal of granulation
tissue, resolution of scar tissue as well as in the restructuring of the basement
membrane around basal keratinocytes (Crawford 1996). It is expressed in the stroma
and has the ability to cleave components of the ECM, such as fibronectin and
vitronectin. Stromelysin-1 can activate several other MMPs including gelatinase
B. Stromleysin-3 is expressed in the connective tissue of developing digits, involuting

mammary glands, and the inflammatory stage of wound healing (Dochtery 1995).

101

Matrilysin

Human matrilysin expression is found in epithelial cells at the sheath of hair
follicles, mammary and parotid glands, pancreas, liver and prostate (Wilson 1996).
In contrast to other stromelysins, matrilysin expression is constitutive, and usually
independent of tissue remodeling. This indicates a function distinct from other
stromal MMPs, such as, proteolytic processing or degradation of peptides in the
luminal space of glandular epithelium. Matrilysin has been found in a large number
of cancers including breast, colon, stomach and prostate. It has also been associated
with invading portions of squamous and basal cell carcinoma of the skin where its
expression has been observed in early and late stages of tumor progression and

shown to affect tumor invasion and tumorigenicity (Sato 1992)

MT-MMPs

A new subfamily of matrix metalloproteinases, the membrane-type matrix
metalloproteinases (MT-MMPs) has been recently described. Membrane type MMPs
(MT-MMPs) have the addition of a transmembrane domain in their protein structure.
The presence of this domain results in their integration into the cell membrane so
that the amino-terminus is positioned outside the cell while the carboxy-terminus is
positioned inside the cell. MT- MMPs are activated by the Golgi serine protease
furin. This is facilitated by the presence of a furin consensus sequence in the domain
structure of MT-MMPs, resulting in their activation before they are expressed at
the cell surface. It has been proposed that the function for MT-MMPs is to activate
gelatinase A at the cell surface in order to have gelatinolytic activity at the leading
edges (invadapodia) of invasive cells. The expression of MT-MMPs is temporally
and spatially regulated, for example, in vascular and urogenital development, and
in the development of osteocartilaginous and musculotendinous structures (Apte
1997). MT-MMPs are also expressed in human placental cells (Takino 1995). These
MMPs are often expressed at the invadapodia of metastatic tumors (Corcoran 1996).

102

EXPRESSION OF MIVIP GENES

The expression of MMP genes is modulated by a number of factors including
cytokines (IL-2, IL-6, and IL-IO), growth factors (TGF-B, TNF-B, and bFGF) and
proto-oncogenes (c-fos, c-jun, and c-ets). The observed distribution of MMPs in
normal cells is, in part, reflected in the promoter elements associated with the genes
that encode them. The signal transmitted by the activated protein kinase (PKC) is
thought to be one of the common mechanisms of regulation in the expression of
several MMP and TIMP genes (Howard 1991). This has been realized through
studies involving the expression of these genes by treatment with the tumor promoter,
TPA (Stearns 1993). The promoters of most human MMPs, except MMP-2 and
Stromelysin-3, contain a TPA responsive element (TRE). This element binds to the
fos/jun-containing complex of the AP-l transcription factor (Sato 1992). This TRE
facilitates the response of MMPs to TPA, IL-l, TNF-a, interferon-[3 (INF-B), and
growth factors, such as, EGF and PDGF. However, the PKC pathway alone cannot
explain the expression pattern of the genes in tumor cells.

All of the MMP promoters, aside from gelatinase A, contain multiple PEA-3
sites. The PEA-3 site was originally identified as an oncogene responsive cis-element
in the polyoma virus enhancer (Pajouh I991). The PEA-3 site binds members of
the c-ets family of nuclear proto-oncogenes, which respond to ras and src activation
(Gutman 1991). This PEA-3 site has been shown to be important for TPA-induced
transcription. Studies have shown cooperativity between AP-l and PEA3 sites
(Gutman 1991). Mutation of either the AP-l site or PEA-3 in the MMP promoter
can reduce TPA and oncogene inducibility by greater than 50%, indicating a
synergistic rather than additive relationship (Matrisian 1992). It is possible that
PEA-3 serves as a fine tuning mechanism for AP-l induction of MMP. Similar

results have been observed in non-MMP enhancers and promoters (Crawford 1996).

103

The expression of Gelatinase A is not inducible by TPA. This is most likely
due to the lack of a TRE in its promoter. Therefore, gelatinase A expression is not
the result of AP-l activity (Murphy 1995). In fact TPA treatment actually inhibits
MMP-2 expression (Corcoran 1996). In addition, the promoter for the gelatinase
A gene also lacks the PEA-3 transcription element (Corcoran 1996). In general,
TGF-B decreases the synthesis of MMP. It has been suggested that this TGF-B
suppression is mediated via a TGF-B-inhibitory element (TIE) on most of the
promoters of MMP genes. The promoter for the gelatinase A gene lacks this TIE.
In fact treatment by TGF-B appears to increase the expression of progelatinase A
in a number of cell types (Corcoran 1996).

Unlike other MMPs, gelatinase A appears to be constitutively expressed in
many cell types (Stearns 1993). Gelatinase A has a non-canonical TATA box and
two SP1 sites (Templeton 1991). Other MMPs have canonical TATA boxes. The
advantage of having a non-canonical TATA box is that it is easier to be recognized
by transcription factors. This sort of promoter is typical of a “house keeping”
gene, and exhibits a rather promiscuous expression pattern, suggesting a role in
cell survival. This sequence has been presented as the basal promoter of gelatinase
A (Templeton 1991).

While general growth-related elements are relevant to MMP expression, it is
important to note that this more restrictive style of promoter lends itself to tight
regulation by a variety of potential promoter and enhancer elements that have yet
to be identified in MMP genes. Therefore, it is expected that elements that control
tissue-specific expression will be identified, in addition to elements that are specific

targets of other signal transduction pathways.

104

MECHANISMS OF MNIP ACTIVATION

The biochemical mechanism for initiating the catalytic activity of mammalian
MMPs is believed to occur via a cysteine switch mechanism (Murphy 1995).
According to this theory, an unpaired cysteine residue in the pro-domain of the
enzyme keeps the zymogen in the latent pro-form by direct coordination with the
zinc atom bound to the catalytic site of MMPs. Four mechanisms for the in vivo
activation of these MMPs have been proposed. The first is by the extracellular
activation of MMPs by proteolytic enzymes. The second is by an autoactivation
mechanism. The third proposed method of activation involves the Golgi network
serine protease furin. Recently, a fourth mechanism of activation has been described,
as the activation of MMP-2 via MT—MMP. Each of these suggested methods of
MMP activation will be explored in relation to degradation of the basement

membrane in cancer invasion.

Activation of MMPs by proteolytic enzymes

Proteases such as, trypsin, plasmin, urokinase, and stromelysin-l can cleave
the prodomain, of most MMPs. The plasmin cascade was one of the first in vivo
mechanisms identified for the activation of MMPs. This mechanism appears to be
involved in the activation of progelatinase B, and prostromelysin-1 and -2.
Plasminogen, when exposed to urokinase-type plasminogen activator (uPA), yields
the active serine protease plasmin (Waghray & Webber 1995). Plasmin can cleave
84 amino acids (the prodomain) from prostromelysin resulting in a fully active
stromelysin (Figure 5.1) (Corcoran 1996). This cascade may occur locally at the
cell surface via the uPA receptor or in the ECM. Active stromelysin will also cleave
the prodomain of MMPs. Stromelysin-l has been observed to activate type I
collagenase and gelatinase B. Recently, matrilysin was reported to activate

stromelysin-l and gelatinase B. The activation of stromelysin-1 by matrilysin was

105

  
  
  

Keratinocytc IL-l \
O -
uPA

Dermal Fibroblast
Pro-interstitial

serum plasminoget\
Collagenase

(52 and 57RDa) uPA

.. ...... . ...... -1

Plasmin N-tenninat cleavage

Pro-stromelysin
(6010a)

Partially Active Active
Interstitial Stromelysin
Collagenase
(42-kDa)
Fully Active I Distant ECM
Interstitial Collagenase Degradation
(40-kDa)

Figure 5.1. Actvation of MMP by plasmin (Corcoran 1996).

106

100% efficient and is similar to plasmin activation of stromelysin. Conversely,
proteolytic activation of pro-gelatinase B by matrilysin is only 50% efficient, this
may be due to the fact that matrilysin cleaves progelatinase B at a site distinct from

the plasmin cleavage site. Gelatinase A and plasmin also activate pro-gelatinase B

(Corcoran 1996).

Autoactivation of MMPs

Autoactivation of MMPs is also possible. The “cysteine switch” mechanism
can be triggered by perturbation of the pro-MMP by certain chemicals, such as,
4-aminophenylmercuric acetate (APMA) (Bu 1995). This disruption initiates a
cascade of events that alters the proMMP protein conformation and allows for the
autoproteolytic cleavage of a portion of the pro-domain, yielding a partially activated
enzyme. Autoactivation of the protease can also be caused by the deletion of the
arginine or cysteine residues in the pro-domain of MMPs. Loss of the amino-terminal
pro-domain is evidence of, but not equivalent to, an active enzyme species. The
autoproteolytic activation of proMMP suggests that the ability to cleave peptide
bonds is acquired prior to removal of the pro segment. Synthetic peptides that
mimic the pro-domain (PRCGXPDV sequence) have been previously shown to inhibit
not only enzymatic activity (Nomura 1993) but also invasion by cancer cells
(Melchiori 1992). Although in vitro studies have aided in understanding some details
for activation of these MMPs, the mechanisms of in vivo activation of these enzymes

is still under investigation.

Activation of MMP by furin
Activation of MMPs by furin occurs prior to secretion of the MMP. Furin is
a Golgi network serine protease that recognizes proteins with a RXKR sequence.

This furin recognition sequence is present in MT-MMPs and stromelysin-3 and is

107

located between the pro-domain and catalytic domain. Cleavage of this domain by
furin results in a fully active form of stromelysin-3 and MT-MMP. Since furin
activates these MMPs in the secretory pathway, their proforms are not normally

present in the extracellular space.

Mechanism for the activation of progelatinase A

While some investigators report that the plasmin cascade is also responsible
for activation of progelatinase A , others have found that plasmin actually degrade
this enzyme. Since cellular activation of this enzyme occurs in the presence of
inhibitors of serine proteases and plasmin, the plasmin cascade may not be the
mechanism utilized in vivo. Unlike other MMPs, the activation of progelatinase A
has been shown to occur at the cell surface. A membrane-associated specific activator
of progelatinase A has recently been described (Imai 1996). The activation of
progelatinase A by membrane preparations from cells treated with these agents can
be blocked by treatment with metalloproteinase inhibitors (including metal chelators
such as EDTA or O-phenanthroline), as well as TIMPs, suggesting that the receptor
involved in this activation may be an MMP.

Membrane Type Metalloproteinase (MT-MMP) has been suggested as the
possible in viva activator of progelatinase A. MT-MMP activates progelatinase A
(Imai 1996). Molecular manipulation of MT-MMP has revealed that its
transmembrane domain is needed for gelatinase activation. When this domain is
deleted or replaced, MT-MMP is soluble and fails to activate progelatinase A (Chen
1996). The working hypothesis for MT-MMP activation of gelatinase A, as
suggested by Corcoran and collaborators (Corcoran 1996), is presented in Figure
5.2. This Figure describes the membrane-associated activation of progelatinase A.
TIMP-2 mediates the binding of pro-gelatinase A to the cell surface by virtue of an

interaction between either MT-MMPs directly or by an unknown separate integral

108

....:..

         
 

Latent complex

. . . ‘ '4 -'1
' .. .‘ f.-,k,. »-‘ '. ~- ‘ "‘ ‘ ' ”'4"
!‘ ~" K-Su ._ "is .‘ "r i " ~ ‘ ' v'r‘r
. 3. . . :.j.~‘ -.~:~.' AT.
(u ‘ aw \. _,-: '~-4 -'4
., , - g -' .' - '- f) »
lnvadopodl
.5144“: i. " , ,- '
n . "f -'.'."" I..'4 x“-
-.. ..-_. },.:- -' 1 x,-
_. ':. " t-.':-:.-..'-.-' "-

V

,3. '-

.. -.'-'< f
in ': v' '5

.'.,' -' ‘ '
.. . 4 .V .
a"... -‘ ' .".4.
~ [4r ' . ...'
an . 1
1* ..
. 5‘

  

.~,~' ”van-3:”. I _' . $4. 3.2.- -'a .1 -...
.. . .' . - u
' . ’ -. - ,2 r

 

TIMP-2

<3

 

Active complex

Cell surface

Gelatinase A

     

 

      
 
 

V tnvadopodia

Gelatinase A

Gelatinase A

 

Inhibited complex 3

Figure 5.2. The activation of gelatinase A by MT-MMP and TIMP-2. The biniding of
TIMP-2 may be mediated by an unknown integral membrane protein (‘2) (Corcoran 1995).

109

membrane protein. Once localized to the cell surface, MT-MMP acts on the
progelatinase A / TIMP-2 complex, and the result is a cell bound activated gelatinase
A / TIMP-2 / MT-MMP complex. When this complex dissociates from the cell
surface TIMP-2 rapidly inactivates gelatinase A. This cell-mediated mechanism
for activation of gelatinase A would favor the degradation of the basement membrane

at the leading edge of invasive cells.

EVIDENCE FOR THE ROLE OF MMPS IN INVASION THOROUGH THE
BASEMENT MEMBRANE

The mechanisms controlling the metastatic progression of a localized tumor
are very complex, and involve many biochemical and cellular events. One likely
event is the expression of MMPs capable of degrading the basement membrane.
Degradation of the basement membrane, which is mainly composed of type IV
collagen and laminin is mediated by a set of secreted MMPs that include gelatinase
A (MMP-2), gelatinase B (MMP-9), Stromelysin-l (MMP-3) and Matrilysin
(MMP-7). Aberrant expression and distribution of these MMPs is a critical part of
the multifactorial invasive nature of tumor cells. The increased expression of MMPs
that degrade the basement membrane has been associated with increasing degree of
invasiveness of tumor cells. In mesenchymal cells, MMP-2 MMP-9 and MMP-3
mRNA transcripts are expressed directly in proportion to their invasive potential
(Sato 1992).

Late stage malignant cell types have been shown to express stromelysins and
interstitial collagenase (Sato 1992). Stromelysin-1 and interstitial collagenase are
not normally expressed in differentiated keratinocytes, however, late stage squamous
cell invasive carcinoma cells express these stromal MMPs. It is important to note
these cancer cell types have acquired a spindle shape and they no longer express
normal epithelial markers like E- and P- cadherins, keratins and desmoplakins and
have begun to express the mesenchymal marker vimentin (Sato 1994).

110

Among the MMPs, stromelysin-1 (MMP-3) possesses the broadest substrate
specificity. Transgenic mice that express an autoactivating mutant of rat MMP-3,
targeted to the epithelial compartment of the mammary gland, have been generated
(Lochter 1997, Sympson 1994). Phenotypically, MMP-3 transgenic mice display
increased branching morphogenesis and lactogenic differentiation at pre-puberty
stages and premature involution during late pregnancy. Branching morphogenesis
requires the invasion of epithelial cells into the adipose tissue, a process reminiscent
of invasion of stromal compartments by tumor cells. Strikingly, a large number of
MMP-3 transgenic mice also develop mammary tumors of various histotypes,
including invasive adenocarcinomas. Studies performed with synthetic inhibitors
of MMP activity and tissue inhibitors of metalloproteinases (TIMPs) have shown
that suppression of MMP activity also suppresses tumor growth and metastasis
(DeClerck 1992). In many cases, the level of MMP-3 expression in tumors of the
mammary gland and other tissues is positively correlated with the degree of
malignancy (McDonnell 1990). An investigation was conducted to determine
whether MMP-3 plays a direct role in invasion of the ECM. Two mammary
carcinoma cell lines, TCLl and SCg6 that formed rapidly growing, invasive tumors
in vivo (Figure 5.3) and migrated through Matrigel and collagen gels in culture,
were used. Antisense oligodeoxynucleotides (ODNs) against MMP-3 inhibited
Matrigel invasion by TCLI and SCg6 cells by more than 80% and collagen invasion
by about 50% (Figure 5.4) (Lochter 1997). These data suggest that MMP-3
expression is associated with invasive properties of these breast cancer cells.

Basement membrane degradation is a key event in tumor invasion and
metastasis. Type IV collagen and laminin are the most abundant of the basement
membrane structural proteins in normal prostate tissue. Selective degradation of
these components of the basement membranes has been observed in the prostate

(Waghary & Webber 1995). A study of the metastatic prostatic adenocarcinoma

111

        
 

    
 
  

‘1?

   

-.. ‘ . H.232?! 5.6% -‘. . _

I ‘Siai’ft—t-h 123.38%}; at"? .-
-_-.-. "-fx': an . .5:
"-1.5333ij 6 38%: 5- “ctr
- .-' ' .4 .‘n‘ "- " J.
X "-3.! orl‘fkdfififik“ '

  

« . 1
-’ ‘ 1.. gingiflgé
'1 "" ’ n 'Q‘f’fj‘g. ( 8

Figure 5.3. Histology of tumors formed by TCLl and SCg6 cells. Hematoxylin/eosin-
stained paraffin sections of primary tumors (panels a and b), invaded skin (panel c) and
lung metastasis (panel (1) generated by TCLl (panels a and c) and SCg6 (panels b and d)
cells injected into the mammary gland of nude mice. The spindle cell tumors (T), entrapped
adipose tissue (asterisks) and epithelial ducts (arrowheads) in the mammary gland are
indicated. Comparison of the regulation of MMP-3 expression by ECM in TCLl and
SCg6 cells with the nonmalignant, functional cell line SCp2 revealed striking differences
that could play a role in the acquisition of an invasive tumor phenotype (Lochter 1997).

112

120 120
3! m
'5 10° = 100
0 0
° 8
h 60 h 60
B o
8 4° 3 4°
3
= 20 g 20
I
o o
N to -' '- I: g: c
n. 5 u 2 O 'E :: .1
U 0 O E g :3 fl. 2
m l- at o 5 g 3. m
o (5 a. 3 3'
a 2 n.
c d
20 m 120
1
O _
= — 1
0 100 8 0°
0
.. an - 8°
0 O
I— .- 6°
3 40 IE 40
E
:3 20 g 20
C
o 0 I
- to to (D (D
1 10 100 g < < <
C :1 n
o " "’
concentration 0 a tn a 6

Figure 5.4. Invasion of Matrigel by SCp2, TCL1, and SCg6 cells and inhibition by proteinase
inhibitors and antisense ODNs. SCp2, TCL1, and SCg6 cells were plated on Matrigel in
modified Boyden chambers. Afier fixation with 5% glutaraldehyde and removal of cells that
had not migrated, the number of cells that had migrated through Matrigel afier 18 h of
culture was quantified by counting the number of toluidine blue-stained cells per visual field.
Panel a, comparison of invasion by SCp2, TCLl and SCg6 cells. Panel b, TCLI cells (black
bars) and SCg6 cells (white bars) were maintained in the absence of proteinase inhibitors
(control) or with GM6001 (10 uM), aprotinin (0.3 uM), leupeptin (1 uM), or pepstatin (1
uM). Panel c, SCg6 cells were maintained in the presence of GM6001 (white circles) or
GM1210 (black circles) at increasing concentrations given in uM. Panel d, TCLl cells (black
bars) and SCg6 cells (white bars) were maintained in the absence of ODNs (control) or in the
presence of SLI antisense (3.2 uM, SL1 AS) or sense (3.2 uM, SL1 S) ODNs, SL3 antisense
ODNs (8 uM, SL3 AS), or collagenase-3 antisense ODNs (32 uM, 03 AS) ODNs. Results
were normalized with control values set to 100 in b-d. Error bars indicate standard deviations
from three experiments. (Lochter 1997)

113

cell line PC-3 showed increased gelatinase A activity. This study indicated that
invasion through the basement membrane was a critical step toward metastasis
(Webber 1995, Webber 1996).

It is important to note that cell invasion is dependent on a cascade of events
that involves the participation of several factors. There are several enzymes capable
of degradation of the basement membrane and other components. The
overexpression and increased activity of MMPs are clearly not the only factors
involved in this process. The turnover of ECM and the expression of MMPs are
also involved in a number of normal conditions. Metastatic invasion is likely to be
due to an imbalance of positive and negative control factors. One such negative
factor may be TIMP. The loss of TIMP expression might also lead to a disregulation
of ECM turnover, giving the same result as MMP overxpression. Results show that
several malignancies have little or no TIMP expression (Murphy 1993). Therefore,

the possibility of treatment of cancer metastasis by TIMP is promising.

CONCLUSIONS

It is now clear that proteases playian integral role in physiological conditions
that require turnover of the extracellular matrix. It has been established that some
MMPs, specifically gelatinases, stromelysins, and matrilysin, will degrade type IV
collagen, a major component of the basement membrane. These MMPs have been
implicated in the invasive behavior of many tumor types. Further study of the
distribution patterns, expression, and activation of these select MMPs will provide
valuable insight into cancer progression. The next chapter will examine the role of

MMPs in prostate cancer.

114

REFERENCES

Apte, S. S., Fukai, N., Beier, D. R., and Olsen, B. R.: The matrix metalloproteinase-14
(MMP- l 4) gene is structurally distinct fi'om other MMP genes and is co-expressed with
the TIMP-2 gene during mouse embryogenesis. J Biol Chem 272: 25511-25517, 1997.

Bu, CH, and Pourmotabbed, T.: Mechanism of activation of human neutrophil gelatinase
B. Discriminating between the role of Ca2+ in activation and catalysis. J Biol Chem 270:
18563-18569, 1995.

Chen, W.: Proteases associated with invadopodia, and their role in degradation of extracellular
matrix. Enzyme Protein 49: 59-71, 1996.

Conway, J .G, Trexler, S.J., Wakefield, J .A., Marron, B.E., Emerson, D.L., Bickefl‘, D.M,
Deaton, D.N., Garrison, D., Elder, M., McElroy, A., Willmott, N., Dockerty, A.J.P., and
McGeehan, GM.: Effect of matrix metalloproteinase inhibitors on tumor growth and
spontaneous metastasis. Clin Exp Metastasis. 14: 115-124, 1996.

Corcoran, M. L., Hewitt, R. E., Kleiner, D. E. Jr, and Stetler Stevenson, W. G: MMP-2:
expression, activation and inhibition. Enzyme Protein 49: 7-19, 1996.

Crawford, H.C., and Matrisian, L.M.: Mechanisms controlling the transcription of matrix

metalloproteinase genes in normal and neoplastic cells. Enzyme Protein, 49: 20-37,
1996.

Crawford, H.C., and Matrisian, L.M.: Tumor and stromal expression of matrix
metalloproteinases and their role in tumor progression. Invasion Metastasis 14: 234-245,
1994-95.

DeClerck, Y. A., Perez, N., Shimada, H., Boone, T. C., Langley, K. E., and Taylor, S. M.:
Inhibition of invasion and metastasis in cells transfected with an inhibitor of
metalloproteinases. Cancer-Res 52: 701-708, 1992.

Docherty, A. J ., O’Connell, J., Crabbe, T., Angal, S., and Murphy, G.: The matrix
metalloproteinases and their natural inhibitors: prospects for treating degenerative tissue
diseases. Trends Biotechnol 10: 200-207, 1992.

Edwards, D.R., Beaudry, P.P., Laing, T.D., Kowal, V., Leco, K.J., ans Lim, MS: The role of
tissue inhibitors of metalloproteinases in tissue remodelling and cell growth. International
J Obesity, 20: S9-SIS, 1996.

Gutrnan, A., Wasylyk, C., and Wasylyk, B.: Cell-specific regulation of oncogene-responsive
sequences of the c-fos promoter. Mol Cell Biol 11: 5381-5387, 1991.

115

Howard, E.W., Builen, EC, and Banda, M.J.: Preferential inhibition of 72- and 92-kDa
gelatinases by tissue inhibitor of metalloproteinases-2. J Biol Chem 266(20): 13070-
13075, 1991.

Imai, K., Ohuchi, E., Aoki, T., Nomura, H., Fujii, Y., Sato, H., Seiki, M., and Okada, Y.:
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,
1 996.

Lochter, A., Galosy, S., Muschler, J ., Freedman, N., Werb, Z., and Bissell, M. J.: Matrix
metalloproteinase stromelysin-l triggers a cascade of molecular alterations that leads
to stable epithelial-to-mesenchymal conversion and a premalignant phenotype in
mammary epithelial cells. J Cell Biol 139: 1861-1872, 1997.

Matrisian LM: Metalloproteinases and their inhibitors in matrix remodeling. Trends Genetics
6(4): 5-15, 1990

Matrisian, L. M.: The matrix-degrading metalloproteinases. Bioessays 14: 455-463, 1992.

McDonnell, S., Navre, M., Coffey, R. J. Jr, and Matrisian, L. M.: Expression and localization
of the matrix metalloproteinase pump-1 (MMP-7) in human gastric and colon
carcinomas. Mol Carcinog 4: 527-533, 1991.

Melchiori, A., Albini, A., Ray, J. M., and Stetler-Stevenson, W. G: Inhibition of tumor cell
invasion by a highly conserved peptide sequence from the matrix metalloproteinase
enzyme prosegrnent. Cancer Res 52: 2353-2356, 1992.

Mignatti, P., and Rifltin, D.B.: Plasminogen activators and matrix metalloproteinases in
angiogenesis. Enzyme Proteins 49: 117-137, 1996.

Murphy, A. N., Unsworth, E. J ., and Stetler Stevenson, W. G: Tissue inhibitor of
metalloproteinases-2 inhibits bFGF-induced human microvascular endothelial cell
proliferation. J Cell Physiol 157: 351-358, 1993.

Murphy, G: Matrix metalloproteinases and their inhibitors. Acta Orthop Scand (Supp 266)
66: 55-60, 1995.

Nomura, S., Hogan, B. L., Wills, A. J ., Heath, J. K., and Edwards, D. R.: Developmental
expression of tissue inhibitor of metalloproteinase (TIMP) RNA. Development 105:
575-583, 1989.

Pajouh, M.S., Nagle, R.B., Breathnach, R., Finch, J .S., Brawer, M.K., and Bowden, GT.:
Expression of metalloproteinase genes in human prostate cancer. J Cancer Res Clin
Oncol I 1 7: 144-150, 1991.

116

Sato, H., Kida, Y., Mai, M., Endo, Y., Sasaki, T., Tanaka, J., and Seiki, M.: Expression of
genes encoding type IV collagen-degrading metalloproteinases and tissue

inhibitors of metalloproteinases in various human tumor cells. Oncogene 7: 77-83,
1992.

Sato, H., Takino, T., Okada, Y., Cao, J., Shinagawa, A., Yamamoto, E., and Seiki, M.: A
matrix metalloproteinase expressed on the surface of invasive tumor cells. Nature 370:
61-65, 1994.

Steams, M.B., and Wang, M.: Type IV collagenase (Mr 72,000) expression in human prostate:
benign and malignant tissue. Cancer Res 53: 878-883, 1993.

Sympson, C. J., Talhouk, R. S.,Alexander, C. M., Chin, J. R., Clifl, S. M., Bissell, M. J ., and
Werb, Z.: Targeted expression of stromelysin-l in mammary gland provides evidence
for a role of proteinases in branching morphogenesis and the requirement for an

intact basement membrane for tissue-specific gene expression. J Cell Biol 125:
681-693, 1994.

Takino, T., Sato, H., Shinagawa, A., and Seiki, M.: Identification of the second membrane-
type matrix metalloproteinase (MT-MMP-Z) gene fi'om a human placenta cDNA
library. the J Biol Chem; 270: 23013-23020 1995.

Templeton, N. S. and Stetler Stevenson, W. G: Identification of a basal promoter for the
human Mr 72,000 type IV collagenase gene and enhanced expression in a highly
metastatic cell line. Cancer Res 51: 6190-6193, 1991.

Waghray, A. and Webber, M. M.: Retinoic acid modulates extracellular urokinase-type

plasminogen activator activity in DU-145 human prostatic carcinoma cells. Clinical
Cancer Res 1: 747-753, 1995.

Wilson, CL, and Matrisian, L.M.: Matrilysin: an epithelial matrix metalloproteinase with
potentially novel functions. Int J Biochem 28: 123-136, 1996.

Webber, M.M., Waghray, A., Bello, Diana, and Rhim, J .S.: Mini review: proteases and invasion
in human prostate epithelial cell lines: implications in prostate cancer prevention and
intervention. Radiation Oncology Invest 3: 358-362, 1996.

Webber, M.M. and Waghray, A.: Urokinase-mediated extracellular matrix degradation by

human prostatic carcinoma cells and its inhibition by retinoic acid. Clinical Cancer Res
1: 755-761, 1995.

117

CHAPTER 6

MATRIX METALLOPROTEINASES AND PROSTATE
CANCER

ABSTRACT

Prostate cancer is the most common cancer currently diagnosed in American
men (Landis 1998). In 1998, 39,200 deaths are expected from this disease, making
prostate cancer second only to lung cancer as the leading cause of death from cancer
among men. Prostate tumors display a high degree of biological variability, ranging
from benign tumors to aggressive adenocarcinomas, and currently there is no method
for determining whether a tumor will remain latent or invade and metastasize. The
biological variability of prostate cancer is measured by the ability of tumor cells to
invade the surrounding basement membrane. This invasion is considered to be the
first step toward metastasis. Matrix metalloproteinases (MMP) are a family a zinc
dependent endopeptidases that degrade components of the extracellular matrix.
MMPs have been strongly implicated in cancer cell invasion and metastasis. This
chapter defines the role of MMPs in prostate cancer by examining the expression
patterns of MMPs and their inhibitors in normal and malignant prostate, especially

in invasive cancer. Further, clinical applications of this work are also explored.

INTRODUCTION

Prostate Carcinoma is one of the leading causes of cancer-related death in
several industrialized countries. In American men, it ranks as the second leading
cause of death from cancer (Powell 1993) after lung cancer. These statistics are
significant because a large segment of the American male population has recently
entered the age group affected by prostate cancer. Prostate cancer is occurs

predominantly in men over the age of fifty. It is estimated that the percentage of

118

men over age 55 will increase from 18% to 25% of the total male population by the
year 2010 (Powell 1993). This increase in the number of affected men will cause
prostate cancer to become a large public health problem.

Prostate cancer displays a large degree of biological variability, ranging from
slow growing to rapidly growing and aggressive metastatic cancer. Distant
metastases to the lymph nodes and bone are the principal causes of death (Festuccia
1995). At the time when prostate adenocarcinoma is first detected it is not currently
possible to determine whether the tumor will remain latent for the patient’s lifetime
or invade and metastasize within 5 years. Despite the existence of several modern
techniques in clinical management of these tumors, little or no progress has been
made in identifying reliable prognostic parameters for predicting the invasive
behavior of an individual tumor. The majority of patients with prostate cancer are
often diagnosed with already disseminated disease (Stephenson 1992). The lack of
good diagnosis and prognostic tools, combined with the relative lack of effective
therapies, makes this disease one of the major causes of cancer related death in men
today. Identification of cellular features that predict the biological behavior of the
individual tumors is one of the major challenges of modern medicine.

The term, invasiveness, describes the ability of cells to cross anatomic barriers
such as basement membranes, interstitial stroma, and intercellular junctions. Normal
invasive processes occur in the embryo and in the adult organism, including
trophoblast implantation, formation of blood vessels or angiogenesis, extravasation
of leukocytes in inflammation, and wound repair. Other invasive processes occur
under pathological conditions, such as bacterial infections and tumor spread.
Invasion is a recurrent theme in the development of malignant tumors. During
tumor growth, while a malignant cell invades normal adjacent tissues, nearby normal
vascular endothelial cells invade the neoplasm itself, by capillaries or venules. These

cells form a tumor’s vascular network (Montironi 1996). A common feature of the

119

invasive process is the degradation of the basement membrane, extracellular matrix
and interstitial stroma (Conway 1996). The ECM-degrading proteinases produced
by the most invasive cells can be divided into three classes, serine proteases,
metalloproteinases, and cystine proteases.

It is apparent that the biological variability of prostate adenocarcinoma
manifests itself in the extent of tumor cell invasion through the epithelial basement
membrane. Destruction of the basement membrane has been observed in many
different cancers including prostate cancer (Conway 1996). Mechanisms controlling
the metastatic spread of a localized tumor are very complex and involve many events.
One such event is the secretion of proteolytic enzymes, capable of degrading the
extracellular matrix by invading tumor cells. Penetration of the epithelial basement
membrane by cancer cells, that can secrete and locally initiate a proteolytic cascade,
is the first step in tumor invasion. The invading cells must also subsequently
penetrate and degrade the basement membrane underlying the vascular endothelium
to enter the circulation (intravasation). The cancer cell must then exit the vascular
network by invading the blood vessel wall of the host target organ (extravasation).
Matrix metalloproteinases and their inhibitors have been proposed as modulators
of these steps in the process of invasion and metastasis. Increased proteolytic activity
of MMPs, specifically gelatinases and stromelysins, has been observed in several
invasive prostate tumor types (Stearns 1993, Wang 1996, Webber 1995). Gelatinases
and stromelysins can use components of the basement membranes as their substrates
and are, therefore, capable of degrading the basement membrane. Studies have
shown a progressive loss of the basement membrane with increasing grade of human
prostate carcinoma (Stearns 1993). These results support the hypothesis that MMP
expression may be indicative of a change from a benign to a malignant state in

prostate cancer.

120

TUMOR CELL INVASION OCCURS IN THREE STEPS

A three-step process has been suggested for describing the sequence of events
during tumor cell invasion and metastasis (Figure 6.1). The first step is the
attachment of the tumor cell via receptors to the ECM glycoproteins, such as,
fibronection and laminin. The next step, following attachment, is the secretion of
proteolytic enzymes by cancer cells and their induction in neighboring stromal cells
and inflammatory cells. These proteases can degrade the components of the
basement membrane. The third step in the process of invasion is the migration of
the tumor cell through the partially degraded ECM. Autocrine motility factors as
well as chemotatic factors can modulate this process. The process of tumor
cell-mediated degradation of the ECM appears to involve a complex series of
proteases that may be secreted and activated in a cascade of activation of latent

proteases (Conway 1996).

EXPRESSION OF MMP IN PROSTATE CANCER

The MMP family is divided into three subclasses based on their homology
and substrate specificity. The three subclasses are the collagenases, the gelatinases,
and the stromelysins. Proteins known as tissue inhibitors of metalloproteinases
(TIMPs) can inhibit the activated form of these MMPs. The process of regulating
tumor cell invasion and metastases may depend, in part, on the balance between
MMPs and their inhibitors. The MMP expression can be considered a positive
influence on invasion while inhibition by TIMP can be considered a negative influence
on invasion. It is thought that the uncontrolled activity of MMPs result in the
degradation of the extracellular matrix and the loss of integrity of the basement

membrane, which facilitates tumor cell dissemination. Disregulation of MMPs has

121

1. Attachment of Tumor Cell

 

 

. tumor cell to the basement membrane
I basement
membrane
2. Secretion of Proteolytic
Enzymes that degrade the
. GA basement membrane
MMP
(A
V
. l' 3. Migration of tumor cell
tirnriior cell through the partially
degraded basement
membrane
O

 

Figure 6.1 Tumor cell invasion occurs in three steps.

122

been observed in many tumor cell types and has been associated with progression
toward metastasis (Stearns 1992). This section will examine the role of MMPs in
metastatic prostate cancer.

Matrilysin (MMP-7) is expressed normally in epithelial cells of the prostate
gland. Its normal function is proteolytic processing of glandular secretions.
Matrilysin can degrades gelatins and fibronectin, and has also been shown to activate
gelatinases and stromelysins. Increased expression of the matrilysin gene has been
associated with progression of adenocarcinoma of the prostate (Pajouh 1991). Using
both Northern blot analysis and in situ hybridization, expression of the matrilysin
gene was observed in 14/ 18 (72%) samples studied (Pajouh 1991). Gelatinase A
was expressed in more than 60% of the sample of prostate adenocarcinomas
examined (Pajouh 1991). In situ hybridization showed that matrilysin was only
expressed in luminal epithelial cells, and not by the surrounding stromal cells (Pajouh
1991). These data suggest that an increase in the expression of matrilysin can lead
to an increase in gelatinase A activation, resulting in a more invasive prostate
adenocarcinoma.

Normal prostate epithelial cells express gelatinase A (Webber 1996). This
expression of gelatinase A has been shown to be upregulated in benign prostatic
hyperplasia and prostate cancer (Stearns 1993). The Gleason score is a measure of
the invasive potential of cancer cells. The tumor is assigned a number 1-10, where
10 represents a highly invasive cancer. Northern blot analysis indicated quantitative
differences in the levels of cytoplasmic gelatinase A (MMP-2) expression between
benign prostatic hyperplasia and prostatic tumors (Figure 6.2). These data indicated
BPH samples showed low expression of gelatinase A, conversely, increasing
expression of gelatinase A was associated with increasing Gleason score in prostatic
carcinomas (Stearns 1993). In situ hybridization showed weak staining for gelatinase

A in BPH, and progressively stronger staining with increasing Gleason score prostate

123

.4
z
>

 

- - - - areas,

 

 

1 2 3 4

5617 a

Figure 6.2. Northern blot analysis of gelatinase A mRNA from prostate tumor tissue (T,
lanes 1-4) and adjacent normal stromal tissue (N, lanes 5-8) respectively. In panel A, the
tissue was from benign prostatic hyperplasia (BPH) tissue (lanes 1 and 5): and from prostate
tumor tissue with a Gleason score (GS) of 1 (lanes 2 and 6); 2 (lanes 3 and 7); or 3 (lanes 4
and 8). In panel B the Gleason score was 4 (lanes 1 and 5); 5 (lanes 2 and 6); 6 (lanes 3 and
7); or 8 (lanes 4 and 8) (Stearns 1993).

124

cancer tumors (Stearns 1993). Stromal tissues surrounding these tumors showed
no expression of gelatinase A. These data associate the increased expression of
gelatinase A with the increasing invasive potential of a cancer cells.

A comparative study of normal prostatic tissue from adults (28-35 years)
and juvenile (4-12 years) patients and diseased prostatic tissue from patients with
benign prostatic hyperplasia (BPH) and prostatic carcinoma (PC) was performed
(Lokeshwar 1993). These studies showed that the secretion of active forms of
MMP was limited to the malignant and juvenile samples (Figure 6.3) (Lokeshwar
1993). The increased expression of MMP seen in these samples indicate a favorable
proteolytic condition for rapid changes in their basement membranes. In the juvenile
prostate, this probably represents normal enlargement of the prostate during the
development of secondary sexual characteristics, while in the adult this change

indicates the onset of BPH or cancer.

 

DActive Gelatinase

>. 8 .Total Gelatinase 8
:- L\TIMPs 2
.d z 7 ‘7 U
13" 4, e 8
< -- I V \
.i 5 \ s a
In In
2:: 4- 4 '
1" 5 Ul
: 3- ‘3 a
v 2- «2 i'.‘
(D

l .1

0 O

 

 

 

Figure 6.3. Examiniation of gelatinase expression, gelatinase activity and TIMP expression
in PC primary carcinoma of the prostate; NAP, normal adult prostate; and JP, juvenile prostate.
*indicates no significant reading (Lokeshwar 1993).

125

The induction of MMP genes can result in the transformation to an invasive
phenotype. DU-145 cells do not express matrilysin (Powell 1993). Powell
transfected DU-145 with a matrilysin cDNA (Powell 1993). These transfected
DU-145 cell overexpressed matrilysin. They were injected into SCID mice to test
their in vivo invasive potential when placed on the mouse diaphragm. The controls
for this experiment were DU-145 cell transfected with vector only. While the control
cells appeared to be inhibited from crossing the diaphragm basement membrane,
DU-145 cells transfected with matrilysin cDNA were able to invade the basement
membrane in vivo (Figure 6.4) (Powell 1993).

Recent studies show an increase in gelatinase A (MMP-2) activity in human
prostate cancer (Wang 1997). This expression of MMP-2 was found to have an
increasing trend corresponding to the Gleason score of the carcinoma. The
expression of MMP-2 in cancers with Gleason scores from 5-8 was low to moderate,
and a noticeable increase was seen in cancers with Gleason score of 9-10. Lymph
node metastases also showed a significant increase in the expression of MMP-2.
This increase was twice as much as the increase observed in Gleason score 9-10
tumors. ELISA measurements, comparing the ratios of activated MMP to proMMP,
showed that this ratio increased as Gleason scores increased (Wang 1997). These
data indicate that both the expression of MMP-2 as well as the activation of MMP-2
are good prognostic measures of the progression of prostatic carcinoma toward a

aggressive phenotype.

MODULATION OF MMP ACTIVITY IN PROSTATE CANCER

The complex process of tumor cell metastasis is thought to be due to several
genetic changes in at least one primary tumor cell subpopulation. These mutations
in tumor cells can yield increased sensitivity to autocrine or paracrine growth factors

or decreased sensitivity to growth inhibitors. Either of these changes results in a

126

 

Figure 6.4. Immunohistochemical staining of SCID mouse diaphragmatic
tumors. DU-l45 cells are shown in panel A, and DU-l45 cells transfected
with matrilysin cDNA are shown in panel B. Panel B shows cells crossing
the basement membrane (arrows) (Powell 1993).

127

selective advantage for the mutated cell subpopulation within both the primary or
secondary (metastatic) tumors. It is possible that a cell with a mutation that results
in‘an increase in MMP activity would result in an invasive or metastatic phenotype.
This section examines the modulation of MMPs by growth factors or growth

inhibitors in the progression of metastatic prostate cancer.

LOSS OF INHIBITION BY TGF-B

Tumor progression to the stage of metastasis may result, in part, from the
selection of certain primary tumor cell clones, which are phenotypically competent
for survival, invasion, and growth at secondary sites. Selection for traits such as
loss of growth inhibitory responses, enhanced motility and increased collagenase
activity are likely to contribute to cancer progression and may be regulated through
the action of growth factors. TGF-B is known to inhibit growth of prostatic epithelial
cells and inhibit most MMPs (Sehgal 1996). Therefore, the elimination or subversion
of TGF-B-responsive pathways should be considered as a mechanistic framework
for metastatic events. One study examined the growth and extracellular matrix
responses to TGF-B in six metastatic and six primary tumor-derived cell lines in a
mouse model of prostate cancer (Sehgal 1996). Tumor cell lines derived from focal
pulmonary metastasis secreted relatively greater quantities of total TGF-B, lost most
or all TGF-B growth inhibition, but responded to TGF-B through induction of
gelatinase B. Cell lines derived from tumors that proliferated at the primary site
retained the growth inhibition and did not produce gelatinase B (Sehgal 1996).
These results suggest that acquisition of differential responses to the TGF-B family
could result in phenotypic traits, which facilitate tumor metastasis from certain
primary site clones. It is conceivable that a mutation occurred in one of the cell
subpopulations of a primary tumor that lead to a metastatic phenotype. This mutated

cell subpopulation was able to invade the basement membrane and enter and exit

128

the circulation to establish a secondary tumor. Any cells collected from these
metastatic tumors would be derived from the mutated cell subpopulation and would

most likely display this similar aggressive phenotype.

MMP SECRETION AND INVASION

Kinesin is a microtubule associated ATPase that powers macromolecule and
organelle transport to the cell surface. Kinesin is ubiquitous and may be universally
important for vesicle transport and secretion in eukaryotic systems. It has been
postulated that enhanced kinesin levels might be a prerequisite for accelerated
protease secretion, as required for cell invasion. Studies examining both invasive
and noninvasive PC-3 human prostate carcinoma cell sublines have demonstrated
interesting results (Stearns 1991). Each of the non-invasive PC-3 sublines failed to
express detectable levels of gelatinase A, while each of the invasive PC-3 sublines
overexpressed gelatinase A. The invasive PC-3 sublines had a significant increase
in the basal expression of kinesin over the non-invasive PC-3 sublines (Stearns 1991).
It is suspected that this increase in kinesin expression results in elevated secretion
of gelatinase A. The elevated kinesin expression observed in invasive PC-3 sublines
can be further increased when treated with conditioned medium collected from any
of the invasive clones. In addition to enhanced kinesin expression, .cells treated
with conditioned medium showed increased gelatinase A secretion and invasion
(Stearns 1991). This coordinately enhanced kinesin expression, gelatinase A
secretion and invasion was not induced by treatment with bovine serum albumin,
fetal calf serum, PDGF, EGF, or IGF. The enhancement of these three factors,
kinesin, gelatinase A and invasion, by conditioned medium, was not observed in
non-invasive PC-3 sublines. These findings suggested the possibility of an autocrine
stimulatory factor as well as the presence of response mechanisms unique to the

invasive PC-3 subpopulations. Trypsinization or treatment with pertussis toxin

129

could inhibit the enhanced response to condition medium (Stearns 1991). Inhibition
by trypsin indicates that the autocrine factor(s) might bind to the surface receptors
of cells to regulate kinesin expression. Inhibition by pertussis toxin suggests that
the receptor may be G-protein-dependent. It is suspected that invasive PC-3 sublines
have acquired the ability to secrete an autocrine growth factor and produce a receptor
to respond to it (Stearns 1991).

Cytokines, such as, interleukin—2 (IL-2), interleukin-6 (IL-6), and interleukin-
10 (IL-10) have been identified as regulators of ECM degradation. Several studies
have demonstrated cytokines as downregulators of MMP expression (Sokoloff 1996,
Stearns 1995, Wang 1996). Currently, the mechanisms and the nature of the second
messenger pathways involved in growth factor and cytokine regulation of MMPs
and TIMPs remain unclear. It is thought that differential regulation of genes that
encode MMP and TIMP may be due to various cytokines and growth factors that
allow responsive elements to either cooperate with, or antagonize each other. Both
IL-2 and Interferon gamma (IFN-y) were demonstrated to down regulate type IV
collagenase in DU-145 and PC-3 human prostate carcinoma cell lines (Sokoloff
1996).

Five different cytokines (IL-2, 4, 6, 10, and IFN-y) and their effects on
MMP-2, MMP-9, TIMP-1 and TIMP-2 expression in HPV-18 immortalized human
prostate carcinoma epithelial cell lines have been recently examined (Wang 1996).
Seven sample immortalized cell lines were derived from three different low-grade
prostatic tumors (Gleason score 5) (HPCA-Sa-HPV18, HPCA-b-HPV18, and
HPCA-c-HPV18) and four high grade prostatic tumors (Gleason score 10)
(HPCA-lOa-HPV18, HPCA-b-HPV18, HPCA-c-HPV18, and HPCA-d-HPV18). All
of the low Gleason score samples (HPCA-Sa-HPV18, HPCA-Sb-HPV18, and
HPCA-Sc-HPV18) showed high levels of expression of TIMP-1 and low levels of
MMP-2 while the high Gleason score samples (HPCA- 1 Oa-HPV18,

130

HPCA-lOb-HPV18, HPCA-lOc-HPV18, and HPCA-lOd-HPV18) showed low
TIMP-1 expression and high MMP-2. ELISA and Northern blot analysis showed
that IL-10 and to a lesser extent IL-4 and IL-6 stimulated significant increases in
the expression of TIMP-1 in all immortalized cell lines, but particularly in the
HPCA-lO-HPV18 group. A decrease in the expression of MMP-2 was also observed
in the HPCA-lO-HPV18 a, b, c cell lines (Figures 6.5 and 6.6) (Wang 1996). The
induction of TIMP-1 in response to IL-6 and IL-10 was found to be both dose and
time dependent (Stearns 1995). These studies purpose that IL-10, IL-6 and IL-4,
in order of effectiveness, can stimulate the expression of TIMP-1 and simultaneously
decrease the levels of MMP-2 expression in epithelial cells derived from the human

prostate.

APPLICATIONS

The expression of MMPs may be used as a biomarker to monitor cancer
progression (Meyers 1996). Prostatic lesions are often identified as atypical (benign)
or dysplastic (possible cancer), based on light microscopic morphological criteria.
The term prostatic intraepithelial neoplasia (PIN) is generally used to describe these
lesions within the ductal and glandular epithelium of the prostate. Frequently
advanced PIN lesions are indistinguishable from malignant prostatic carcinoma only
on the basis of cell morphology. This similarity may suggest a premalignant role of
PIN. A comparison of both phenotypic and genotypic characteristics of PIN versus
prostatic cancer cells would be useful in the elucidation of this likely transformation.
As a result of the focal nature of PIN, studies of biomarkers such as growth factor
receptors, oncogene products, and glycoslyated tumor antigens are difficult, because
it is almost impossible to isolate pure preparations of PIN. One technique that has
been used is micro-dissection. This procedure is tedious and may still yield tissue

samples contaminated by surrounding stroma or normal epithelial cells. This

131

 

 

 

 

 

 

 

[:1 Series 1 700
Series 2
I Series 3 500 H L
D Series 4 A 500
E
g 400
N
d 300
2
2 i
200
100
0

 

 

 

 

 

 

 

 

 

 

 

 

0 .1 1 5 10
Cytokine (ng/ml)

Figure 6.5. Concentration-dependent effect of Bar series (1) IL-lO, (2) IL-6, (3) IL-2, and

IL-4 on the production of MMP-1 by HPCA-lOaI-IPVIS prostate cancer cell line (Wang
1996)

 

 

 

 

 

 

 

 

 

 

C] Series 1 700
I Series 2
I Series 3 600
D Series 4 a 500
E
g 400
E 300
E
*- 200
100
0 A A +1 e. m
0 .1 1 5 1O
Cytokine (ng/ml)

Figure 6.6. Concentration—dependent effect of Bar series (1) IL-lO, (2) IL-6, (3) IL-2, and

(4) IL-4 on the production of TIMP-1 by HPCA-lOaHPV18 prostate cancer cell line (Wang
1996)

132

contamination makes differentiation of biomarker expression between the basal
versus luminal components of the gland or duct impossible. Therefore, an
immunohistochemical technique combined with in situ hybridization has been
suggested (Myers 1996). A recent study used MMP expression to clarify the
relationship of PIN to prostatic adenocaricnoma (Myers 1996). The expression of
MMP was observed in PIN cells, normal cells, and malignant cells. Enhanced MMP
activity was associated with dysplastic PIN cells and malignant cells. This might
represent an important event in the development of prostatic adenocarcinoma. It
has, therefore, been suggested that the observation of MMP expression via in situ
hybridization studies may serve as a useful biomarker for prostate adenocarcinoma
progression (Myers 1996).

An increase in MMP-2 (gelatinase A) immounostaining was recently
demonstrated in the progression of prostate cancer (Montironi 1996). This study
showed little or no staining in normal prostate, moderate staining in PIN and heavy
staining in invasive prostatic adenocarcinoma. This expression of MMP-2 appeared
to be localized to the cells in contact with the stroma or close to the stroma (Figure
6.7, 6.8, & 6.9) (Montironi 1996). These findings further support the use of MMPs
as biomarkers for malignant conversion.

Taxol has been used to block MMP-2 (gelatinase A) secretion in PC-3 cell
lines (Stearns 1992). Taxol is an alkaloid with great potential in therapeutic
treatment of cancer. The main effect of taxol is stabilization of microtubule
structures. Taxol binds tubulin to form abnormally stable taxol microtubules this
action disrupts the dynamic state required for normal function in the cell (Figure
6.10) (Stearns 1992). It was shown that taxol blocks protease secretion and a dose
dependent block of invasion and metastasis in vitro and in viva occurred. These
results suggest that taxol has a therapeutic potential for the inhibition of invasion,

by its ability to block MMP-2 secretion (Stearns 1992). Mechanistically, targeted

133

        

       
   
   

g. 'l . 6‘ “\J ..
1 q . . ~
, “:33an .5 .fi‘,‘ .
5'51. . :.?‘:1.? a. “

i '5‘
.fm:2a;j:.3m 11, , “f 4
.3" ‘ '

Figure
by arrows) for MMP-2 in the basal cells and little immunoreactivity in the secretory
epithelium (Montironi 1996).

V4 .9; ~.- ’
"*‘é‘a'az‘qa‘agfl
’ '-c -;'l‘-

.. in; ”2;..- = :o::‘ _. 4 -
Figure 6.7. High grade prostatic intraepitheial neoplasia shows intense MMP-2
staining (indicated by arrows) in cells mainly in the cell layer adj cent to the stroma
(Montironi 1996).

134

E
I-

”L

 

Figure 6.8. Prostatic adenocarcinoma shows intense immunostaining for
MMP-2 (arrows) in small clusters of neoplastic cells located in the periphery
of the tumor nodules at the level of advancing front of the tumor (Montironi
1996)

   

Figure 6.9. Immunofluorescence images of PC-3 ML cells labeled with
beta-tubulin anti bodies and a secondary antibodiy conjugated to fluoresein
isothiocyanate. Panel A, show untreated cells. Panel B, shows PC-3 ML cells
treated with 0.5 mM taxol for 6 hour, the more intense signal is due to the
stabilization of microtubule bundles (arrows). This stabilization had an
inhibitory effect on MMP-2 secretion (Stearns 1992).

135

chemotherapy, distinct from cytotoxic agents that can interfere with prolonged
invasion, could prove valuable in the treatment of established prostate
adenocarcinoma.

All-trans Retinoic Acid (ATRA) has been shown to be effective in the
inhibition of uPA by increasing PAI-l levels (Kim 1995, Waghray & Webber 1995)
. This effect is similar to the blockage of gelatinase A (MMP-2) and gelatinase B
(MMP-9) in that it results in the inhibition of ECM degradation (Waghray & Webber
1995). Plasmin is involved in the degradation of the basement membrane and has
been implicated in a proteolytic cascade that result in the activation of several MMPs
and the subsequent breakdown of the basement membrane (Kim 1995, Webber 1996).
Treatment with ATRA resulted in a decrease in the net activity of uPA, gelatinase A
and gelatinase B (Waghary & Webber 1995). Although N-(4-Hydroxyphenyl)-
retinamide (4-HPR) another retinoid, increased uPA expression, the net uPA activity
was decreased because of a marked increase PAI-l expression (Kim 1995). Both
treatment with ATRA and treatment with 4-HPR decreased invasion and increased
in PAI-l (Kim 1995, Waghray & Webber 1995, Webber 1996).

Dahiya showed that 13-cis Retinoic Acid (RA) blocked invasion by the human
prostatic carcinoma LNCaP cell line (Dahiya 1994). This blockade might occur in
two ways: first 13-cis RA significantly inhibits the binding of LNCaP cells to the
basement membrane and second l3-cis RA inhibits the degradation of type IV
collagen. The mechanism was not fully characterized but thought to be caused by
the ability if 13-cis RA to modulate the expression of various genes both in vitro
and in vivo. These data suggest that retinoids have the potential for use as
chemopreventive and chemotherapeutic agents, however, their modulation of
invasion is not fully understood. Therefore, more research is indicated.

Cinnamic acid is a naturally occurring fatty acid. This compound is widely

distributed in the plant world and is commonly use to add flavor in cooking. This

136

substance appears to have an antitumor effect against a broad spectrum of human
tumors at doses that have no significant effect on normal cells. In a recent study,
PC-3, DU-145, and LNCaP prostate cancer cell lines were tested in response to
treatment with cinnamic acid. A significant decline in cell growth was observed in
all of these cell lines (Liu 1995). These studies showed a decrease in the expression
of gelatinase A, as well as an increased expression of TIMP-2 in human melanoma
cells (Liu 1995). It is possible that cinnamic acid might also serve as a possible

chemopreventive agent in the treatment of prostate cancer.

CONCLUSIONS

Prostate cancer is the second leading cause of death from cancer in American
men. Several steps in tumor growth and metastasis require proteolytic degradation
of the extracellular matrix. This ECM degradation and subsequent invasion is a
recurrent theme throughout the process of cancer metastasis. MMPs and TIMPs
are major players in these processes. MMPs and their inhibitors in the normal cell
aid during development and wound healing. In contrast, increased MMP expression
in transformed cells can lead to progression toward cancer invasion and metastasis.
MMP overexpression has been identified in highly invasive prostatic cancer cell
lines, and induced MMP overexpression has been shown to increase the invasive
potential of non-invasive prostatic cell lines.

The saving grace to this grim picture is the fact that the transformation of
normal prostate cells to malignant adenocarcinoma occurs over a long latent period
of 20-30 years. This period of latency provides a window of opportunity for both
treatment and prevention. With further understanding of the role of MMP in prostatic
malignant transformation, important chemotherapeutic and chemopreventive agents

can be developed to combat this disease.

137

REFERENCES

Conway, J. G, Trexler, S. J ., Wakefield, J. A., Matron, B. E., Emerson, D. L., Bickett, D.
M., Deaton, D. N., Garrison, D., Elder, M., McElroy, A., Willmott, N., Dockerty, A.
J ., and McGeehan, G. M.: Effect of matrix metalloproteinase inhibitors on tumor
growth and spontaneous metastasis. Clin Exp Metastasis 14: 115-124, 1996.

Dahiya, R., Park, H. D., Cusick, J ., Vessella, R. L., Fournier, G, and Narayan, P.: Inhibition
of tumorigenic potential and prostate-specific antigen expression in LNCaP human
prostate cancer cell line by 13-cis—retinoic acid. Int J Cancer 59: 126-132, 1994.

Festuccia, C., Vincentini, C., di Pasquale, A. B., Aceto, G, Zazzeroni, F., Miano, L., and
Bologna, M.: Plasminogen activator activities in short-term tissue cultures of benign
prostatic hyperplasia and prostatic carcinoma. Oncol Res 7: 131-138, 1995.

Kim, J. H., Tanabe, T., Chodak, G W., and Rukstalis, D. B.: In vitro anti-invasive effects of
N-(4-hydroxyphenyl)-retinamide on human prostatic adenocarcinoma. Anticancer
Res 15: 1429-1434, 1995.

Landis, S. H., Murray, T., Bolden, S., and Wingo, P. A.: Cancer statistics, 1998. CA Cancer
J Clin 48: 6-29, 1998.

Liu, L., Hudgins, W. R., Shack, S., Yin, M. Q., and Samid, D.: Cinnamic acid: a natural
product with potential use in cancer intervention. Int J Cancer 62: 345-350, 1995.

Lokeshwar, B. L., Selzer, M. G, Block, N. L., and Gunja Smith, Z.: Secretion of matrix
metalloproteinases and their inhibitors (tissue inhibitor of metalloproteinases) by human
prostate in explant cultures: reduced tissue inhibitor of metalloproteinase secretion
by malignant tissues. Cancer Res 53: 4493-4498, 1993.

Montironi, R., Lucarini, G., Castaldini, C., Galluzzi, C. M., Biagini, G., and Fabris, G.:
Immunohistochemical evaluation of type IV collagenase (72-kd metalloproteinase)
in prostatic intraepithelial neoplasia. Anticancer Res 16: 2057-2062, 1996.

Myers, R. B. and Grizzle, W. E.: Biomarker expression in prostatic intraepithelial neoplasia.
Eur Urol 30: 153-166, 1996.

Pajouh, M. S., Nagle, R. B., Breathnach, R., Finch, J. S., Brawer, M. K., and Bowden, G T.:
Expression of metalloproteinase genes in human prostate cancer. J Cancer Res Clin
Oncol 117: 144-150, 1991.

Powell, W. C., Knox, J. D., Navre, M., Grogan, T. M., Kittelson, J., Nagle, R. B., and
Bowden, G T.: Expression of the metalloproteinase matrilysin in DU-145 cells

increases their invasive potential in severe combined immunodeficient mice. Cancer
Res 53: 417-422, 1993.

138

Sehgal, I., Baley, P. A., and Thompson, T. C.: Transforming growth factor betal stimulates
contrasting responses in metastatic versus primary mouse prostate cancer-derived
cell lines in vitro. Cancer Res 56: 3359-3365, 1996.

Sokoloff, M. H., Tso, C. L., Kaboo, R., Taneja, S., Pang, S., deKernion, J. B., and Belldegrun,
A. S.: In vitro modulation of tumor progression-associated properties of hormone
refractory prostate carcinoma cell lines by cytokines. Cancer 77: 1862-1872, 1996.

Stearns, M. E. and Wang, M.: Regulation of kinesin expression and type IV collagenase
secretion in invasive human prostate PC-3 tumor sublines. Cancer Res 51: 5866-5875,
1991.

Stearns, M. E. and Wang, M.: Taxol blocks processes essential for prostate tumor cell (PC-3
ML) invasion and metastases. Cancer Res 52: 3776-3781, 1992.

Stearns, M. E. and Wang, M.: Type IV collagenase (M(r) 72,000) expression in human
prostate: benign and malignant tissue. Cancer Res 53: 878-883, 1993.

Stearns, M. E., Wang, M., and Stearns, M.: IL-lO blocks collagen IV invasion by “invasion
stimulating factor” activated PC-3 ML cells: upregulation of TIMP-1 expression.
Oncol Res 7: 157-163, 1995.

Stephenson, R. A., Dinney, C. P., Gohji, K., Ordonez, N. G, Killion, J. J ., and Fidler, I. J .:
Metastatic model for human prostate cancer using orthotopic implantation in nude
mice. J Natl Cancer Inst 84: 951-957, 1992.

Waghary, A. and Webber, M. M.: Retinoic acid modulates extracellular urokinase-type
plasminogen activator activity in DU-145 human prostate carinoma cells. Clin Cancer
Res 1: 747-753, 1995.

Wang, M., Fudge, K., Rhim, J. S., and Stearns, M. E.: Cytokine regulation of the matrix
metalloproteinases and their inhibitors in human papillomavirus-18 transformed human
prostatic tumor cell lines. Oncol Res 8: 303-15, 1996.

Webber, M. M. and Waghray A.: Urokinase-mediated extracellular matrix degradation by
human prostate carcinoma cells and its inhibition by retinoic acid . Clin Cancer Res 1 :
755-761, 1995.

Webber, M. M., Waghray A., Bello, D., and Rhim, J. M.: Mini review: proteases and invasion
in human prostate epithelial cell lines; implications in prostate cancer prevention and
intervention. Radiation Oncol Invest 3: 358-362, 1996.

139

CHAPTER 7

COMMON MECHANISMS OF
MATRIX METALLOPROTEINASES
IN NORMAL AND PATHOLOGICAL CONDITIONS
INVOLVING EXTRACELLULAR MATRIX TURNOVER

ABSTRACT

Several normal and pathological conditions involve the turnover of
extracellular matrix. Conditions such as wound healing, regeneration, endometrial
breakdown, mammary gland involution, ductal morphogenesis, angiogenesis and
apoptosis are normal conditions involving ECM turnover. Major pathological
conditions such as cancer invasion and metastasis, chronic wounds, and arthritis,
are conditions that require ECM degradation. Previous chapters have identified
matrix metalloproteinases as important factors involved in ECM turnover. Since
the pioneering work of Gross (Gross 1962) in the identification of collagenase type
I, much research has been conducted in an effort to determine the role of MMPs in
other conditions characterized by ECM turnover. This chapter highlights some of
the common mechanisms of MMP expression and secretion, activation and MMP
inhibition as described in previous chapters and compares and contrasts them in

normal and pathological conditions.

INTRODUCTION

The extracellular matrix is a polyfunctional mosaic composed of various
molecules such as collagens, elastin, proteoglycan, fibronectin and laminin. This
intercellular medium affects fundamental cell functions such as cell adhesion,
proliferation and differentiation. The early notion that ECM is an inert and stable
structure has been dispelled and it is now clear that a dynamic equilibrium between

synthesis and degradation of ECM components is required for matrix maintenance.

140

The active destruction of ECM integrity by members of the matrix metalloproteinase
family occurs during processes as diverse as development, wound healing,
regeneration, arthritis and cancer metastasis. Since several of these normal and
pathological processes require the turnover of extracellular matrix, and MMPs
degrade the ECM, it can be deduced that MMP expression in these process is a
common mechanism of ECM turnover.

This chapter seeks to identify some common mechanisms involved in ECM
turnover by MMPs. It will use development, wound healing and regeneration as
models of normal conditions, and arthritis, glomerular nephropathy, myocardial
infarction and cancer as models of pathological conditions. By comparing and
contrasting regulation versus disregulation of MMPs in these normal and pathological
models this review provides valuable information. This information can be applied

to further research into the treatment of several life-threatening diseases.

NORMAL CONDITIONS

Matrix metalloproteinases were first described in Xenopus Leavous, the
African Clawed Frog. Interstitial collagenase (MMP-1) was identified as the active
protein involved in the degradation of collagen in tadpole tail tissue explants (Gross
1962). The protein structure of interstitial collagenase was subsequently accepted
as the prototypic member of the family of MMPs (Goldberg 1992). In addition to
development, there are other normal conditions that are also characterized by ECM

turnover.

Development
It has been speculated that normal cell proliferation, and migration of
fibroblasts in vivo require the selective secretion of matrix degrading enzymes. It

is suspected that MMPs and their inhibitors can modulate these processes through

141

their role in ECM remodeling. Expression of MMPs can affect the release of a cell
from its normal matrix interactions to allow for growth and division. MMPs can
also facilitate directional degradation required for the migration of daughter cells
observed in development.

During amphibian metamorphosis, the intestinal epithelium dramatically
changes due to apoptosis of larval cells and the proliferation and differentiation of
adult cells. This transition from larval to adult epithelium appears to be dependent
on the remodeling of the ECM and is induced by thyroid hormone (TH). In one
study, the TH response genes from Xenopus small intestinal cells were cloned in the
efforts to identify the molecular mechanisms of this induction. One of these early
response genes is the Xenopus homolog of the mammalian stromelysin-3 (MMP-11)
gene (Ishizuya-Oka 1996). Stromelysin-3 has recently been described, and like
other MMPs, it has a prodomain, a zinc binding catalytic domain, and a
hemopexin-like binding domain. However stromelysin-3 differs markedly from
previously described MMPs in that has a furin recognition sequence and can be
activated by the Golgi enzyme furin in the secretory pathway and released as a 45
kDa active enzyme. The substrates for stromelysin-3 are also distinct in that
stromelysin-3 does not appear to cleave the classic MMP substrates such as gelatin,
casein, and elastin. Stromelysin-3 does cleave serine protease inhibitors,
(11-proteinase inhibitor and a2-antitplasmin. The degradation of these substrates
by stromelysin-3 can facilitate an increase in elastase and plasmin activity. The
expression of stromelysin-3 has been associated with epithelial regression as seen
in mammary glad involution (Noel 1996), limb bud interdigitation and breast tumor
cell invasion (Basset 1993).

A study using in situ hybridization was performed to examine the expression
of the stromelysin-3 gene in response to TH in the development of Xenopus small

intestine (Ishizuya-Oka 1996). in situ hybridization is a technique that uses a labeled

142

anti-sense RNA probe to detect the presence or absence of a particular mRNA, in
this case the MMP-ll mRNA. The results of these in situ hybridization studies are
summarized in Figure 7.1 (Ishizuya-Oka 1996). The small intestines of Xenopus
tadpoles were examined from stage-55 (four weeks after fertilization) to the stage-66
(end of metamorphosis). There are three layers in the small intestine of Xenopus,
beginning from the lumen; they are the epithelial layer, the connective tissue layer,
and the muscular layer. In the larval intestine there is a single fold into the lumenal
space, called typhlosole (Ty). The thickest region of connective tissue is located in
this region. As normal development proceeds from a larval epithelium, to an adult
epithelium more folds are developed leading to a more absorptive organ. In the
early stages (55-59), stromelysin-3 mRNA is detected only in a few cells in the
connective tissue layer and only in the lower regions near the muscular layer of the
small intestine (Figure 7.1 panel A and B). At stage-61 most of the cells in the
connective tissue layer stained for stromelysin-3 mRNA, several of these cells are
right under the epithelial layer. The increase in the expression of stromelysin-3
coincides not only with a remodeling of the basement membrane but also the
degeneration of larval epithelium and the rapid growth of adult epithelial primordial
cells (Figure 7.1 panel C). Once the larval epithelium is replaced by the adult
epithelium at stage-65, no stromelysin-3 mRNA positive cells are detected (Figure
7.1 panel D). The transitory staining for stromelysin-3 mRNA may represent the
migration of stromal elements expressing stromelysin-3 from the muscular layer
toward the epithelial layer. These results suggest that the remodeling of the basement
membrane, induced by stromelysin-3, promotes the epithelial transformation from

larval to adult by inducing regional apoptosis and selective cell survival.

143

 

Figure 7.1. Cross-sections of the anterior part of the small intestine hybridized with the
antisense stromelysin probe. The small intestine at stage 59 is shown in panel a. The layer of
connective tissue (CT) is thin except for the typhlosole (Ty). Signals (arrows) are observed
in some cellsof the connective tissue near the muscular layer (M), but are weaker in the
upper region of the typhlosole x150. Panel b shows a higher magnification of the bottom
region of the typhosole at stage 59. Most of the cells in the connective tissue show a positive
reaction x610. The small intstine at stage 61 is shown in panel c. The basal surface of the
epithelial primordia (astrisks) into the connective tissue. Most of the connective tissue cells
just beneath the epitheilum are positive (arrows) x610. Intestinal fold at stage 65. No
positive cells are seen in the connective tissue cells x490. Bars:20 um. LE larval epithelium;
AE adult epithelium; L lumen (Ishizuya-Oka 1996).

The co-expression of membrane-type matrix metalloproteinases (MT-MMP)
and tissue inhibitor of metalloproteinase-2 (TIMP-2) has been studied in mouse
development. MT-MMP has been shown to activate gelatinase A (MMP-2) at the
surfaces of cells (Lee 1997, Sato 1994). This proposed molecular model for
activation of progelatinase A is likely to be most relevant in situ at locations where
MT-MMP and TIMP-2 are produced by the same or adjacent cell populations. Such
is the case in mouse development. The co-expression of MT-MMP and TIMP-2
genes during mouse development suggests a possible mechanism for the activation
of gelatinase A.

Recent work from a number of laboratories have resulted in the discovery of
a subfamily of MMPs, the membrane-type or MT-MMPs (Sato 1994, Takino 1995).
These MT-MMPs are distinct from other MMPs in that they are type I membrane
proteins (N terminus of catalytic domain is extracellular) and have a single
membrane-spanning region. MT-MMP has the potential to be activated by furin or
furin-like enzymes. The activation complex is trimolecular and involves MT-MMP,
progelatinase A and TIMP-2. A working theory of gelatinase A activation by
MT-MMP and TIMP is diagramed in Figure 7.2. TIMP-2 and MT-MMP interact at
their NH2 terminal domains (Imai 1996) (Figure 7.2 panel A). The interaction of
MT-MMP and TIMP-2 allows each of these molecules to interact with gelatinase A
(Figure 7.2 panel B). Once the three molecules come together MT-MMP activates
proMMP-2 by cleaving the pro-domain. Gelatinase A is active as long as the
trimolecular complex remains intact, when the complex is broken MMP-2 is
immediately inactivated by TIMP-2. Since activation of gelatinase A by MT-MMPs
appears to be facilitated by TIMP-2, the co-expression of TIMP-2 and MT-MMP

would provide the necessary environment for gelatinase A activity.

145

 

Figure 7.2. MMP-2 Activation by MT-MMP/I‘ IMP complexes with MT-MMP (panel A),
pro MMP-2 associates with the MT-MMP/TIMP complex and is activated by MT-MMP
(panel B), The MMP-2/1'IMP/MT -MMP complex is active as long as the trimolecular complex
is together (panel C). Once MMP-2 breaks free it is immediately inactivated by TIMP.

146

Using in situ hybridization antisense probes for MTl-MMP and TIMP-2,
the co-expression of MT-MMP and TIMP-2 genes in developing mice was observed
(Figure 7.3 through 7.6). MT-MMP and TIMP-2 were prominently expressed in
large arteries like the aorta and umbilical arteries (Figure 7.3). In addition, low
levels of expression were seen in the myocardium, with more prominent expression
in the cardiac outflow tract as it approaches the root of the aorta. The expression
appears in smooth muscles of the developing tunica media. Co-expression of
MT—MMP and TIMP-2 during bladder development and in specific urogenital organs
further demonstrated this association with smooth muscle (Figure 7.4). This
co-expression of MT-MMP and TIMP-2, in the areas that comprise smooth muscle
cells in these organs may be important for smooth muscle cell migration.

Prominent co-expression of MT-MMP and TIMP-2 was seen in tendons,
ligaments, muscle, and joint capsules (Figure 7.5). The ECM of these tissues is
principally composed of fibrillar collagens, in particular, types I and III collagens.
A recent report demonstrated that MT-MMP has the ability to cleave collagen and
gelatin (Ohuchi 1997). Together, these observations suggest that MT-MMP may
be involVed in the remodeling of dense connective tissues, not just as an activator
of proMMP-2 but as a matrix degrading enzyme in its own right, with the potential
to deliver both collagenolytic and gelatinolytic activity. This finding may be relevant
to the etiology of joint contractures and the repair of ligaments and tendons.

Expression of MMP-14 and TIMP-2 has also been observed at the placental
implantation site (Apte 1997). At 12.5 and 14.5 days post coitum, these transcripts
are co-localized to the junctional region between the uterine deciduum and placental
spongiotrophoblast with minimal labeling in the labyrinth zone of the placenta (Figure
7.6). This expression of MMP may facilitate cellular invasion observed during
implantation. The expression of MMP has been associated with several other

reproductive mechanisms.

147

 

Figure 7.3. Mmpl4 (A, D, and E) and Timp2 (B) are expressed in the media of muscular
arteries. A-C, show the cardiac outflow tract at 12.5 days p.c. Note hybridization of the
aorta (a) but not of the heart (h) in panels A (Mmp14) and B (Timp2). C, shows the cardiac
outflow region hybridized to the Mmpl4 sense probe. D, abdominal aorta (a) at 12.5 days
p.c. hybridized to Mmpl4 antisense probe. Note labeling in the media (see enlarged box
alongside at upper right) and the absence of labeled cells in the endothelial layer. v, vertebral
column; m, posterior abdominal mesenchyme. E, umbilical artery at 17.5 days p.c. hybridized
to Mmpl4 antisense probe. Boxed area is enlarged alongside (shown at lower left of figure)
showing that labeling is restricted to the media. Arrowheads in panels D and E indicate the
endothelium (Apte 1997).

148

 

Figure 7.4. Hybridization of Mmpl4 probe during urogenital development. A, transverse
section through the developing urinary bladder (b) at 14.5 days p.c. showing labeled cells in
the muscular wall (the lumen of the urogenital sinus is not seen in this section). Note also
labeling of umbilical arteries (a) B, labeled cells are seen in and around the bulbourethral
gland (b) adjacent to the bulbar urethra (u). C and D, expression of Timp2 in ganglia of the
17.5 day mouse embryo. C, dorsal root ganglia (g); D, trigeminal ganglion (Apte 1997).

149

 

Figure 7.5. MMP-14 and TIMP-2 are expressed in developing synovial joints. Shown are
knee joints of 17.5 days p.c. embryos hybridized to an Mmpl4 (A) or Timp2 (B) antisense
probe. Hybridization is seen with cells of the posterior cruciate ligament (PCL), anterior
cruciate ligament (ACL), and patellar tendon (PT). Note labeling of cells in superficial layers
of articular cartilage in both femur (f) and tibia (t) and the absence of labeled cells in cartilage
elsewhere (Apte 1997).

 

Figure 7.6. Expression of Mmpl4 (A) and Timp2 (B) at the placental implantation site (12.5
days p.c.). Similar regions at the periphery of the implantation site are shown. m, myometrium;
p, placenta (labyrinthine portion). The labeling appears in the junctional area that includes
the decidual reaction of the uterus and spongiotrophoblast of the placenta (Apte 1997).

150

Endometrial cycling

In normal ovulation cycles the human endometrium undergoes complex
changes, involving the proliferation and differentiation of both epithelial and stromal
components, in preparation for embryo implantation. In a non-conception cycle,
circulating estradiol and progesterone levels decline, in response to the regression
of the corpus luteum. This initiates a complex process that leads to the breakdown
and shedding of endometrial lining. Traditionally, it has been assumed that this
process is a result of the ischemic necrosis caused by vasospasm of the endometrial
spiral arteries. There is strong evidence to support the role of proteases in this
process.

Studies have shown that MMP expression is responsive to progesterone
withdrawal. Progesterone withdrawal is a physiological stimulus for menstruation.
MMP-2 (gelatinase A) levels increase in stromal cultures subsequent to progesterone
withdrawal and coincident to an upsurge in MMP activity. There is evidence to
suggest that steroid hormones modulate MMP-2 expression (Figure 7.7 panel A)
(Irwin 1997). Stromal endometrial cells were treated with 10nM estradiol and 111M
progesterone, followed by one of three treatments: 1) total steroid withdrawal; 2)
treatment with 0.2 mM progesterone; 3) treatment with 10 nM estradiol, 0.2mM
progesterone, and 1 mM RU486 (progesterone antagonist). The expression of
MMP-2 increased with total steroid withdrawal and With the estadiol, progesterone
and RU486 treatment. The increased gelatinase A activity observed in these
treatments was not detected in cells that received the progesterone treatment alone.
Integrative densitometery showed a six-fold increase in gelatinase expression after
4 days withdrawal of progesterone (Figure 7.7 panel B) this was followed by a
decrease in expression after 6 days. These data indicate that the withdrawal of
progesterone is an important modulator in the expression of MMP-2 for the turnover

of ECM observed in endometrial cycling.

151

MWabcdef

97-
68' - -

43-

w sr Ei-z Hec
as

1A
B
MW 3 b c d e f g II I
97-
68. - - . ——— -- —~
43-
2_L_£ 2.4—6. 2.4—6. <-days
P EP W
+RU486
C
300-
‘1'
E 200-
E.
5 100-
o.

 

 

2.4—6
P EP+RU486 W

- Figure 7.7. Western irnmunoblot analysis of MMP-2 in stromal cell cultures. Samples of 600
pl (A) or 400 pl (B) of 2 d-conditioned media were concentrated, electrophoresed under
nonreducing (A) or reducing (B) conditions, and analyzed by Western immunoblotting with
anti-WIP-Z. Position of the molecular weight markers (MW) is indicated on the lefi margin.
(A) Stromal cells (ES) were cultured 14 d in serum-flee medium as described in Fig. l, in the
absence of steroids (C, lane a), in the presence of 10 nM E + 1 uM P (EP, lane b), or received
10 nM E + 1 uM P for 14 d followed by steroid withdrawal for 4 d (W, lane c). Human skin
fibroblasts (SF, lane (1), human endometrial epithelial cells (EE, lane e), and HEC-IA cells
(lane 1), received no steroids. (B) Stromal cells were cultured 14 d inserum-free medium
containing 10 nM E +1 uM P followed by 6 d in the presence ofO.2 uM P (P), 10 nM E +
0.2 uM P + 1 uM RU486 (EP + RU486), or nosteroids (W), and conditioned media collected
2, 4, and 6 d afier the time of withdrawal. (C) Integrative densitometry of the 72-kD MMP-
2 band on Western blots of 2 d-conditioned media collected 2, 4, and 6 d alter the time of
withdrawal as described for B. Levels of immunoreactive MMP-2 (IR.MMP-Z) in
conditionedmedia were estimated running in parallel known amounts (1 0-500 ng) of purified
MMP-2. Values are the meaniSEM (bars) of measurements on two blots (Irwin 1997).

152

A temporal expression of MMP-2 (gelatinase A) was observed in-the human
endometrium during the menstrual cycle. Northern blot analysis was performed to
examine the expression of MMP-2 (Figure 7.8 panels A and B). The expression of
both MMP-2 and B actin were examined, $-actin expression remains consistent
throughout the proliferative and secretory phases, and is used as an internal standard
to assess the relative amount of mRNA loaded into each lane. The MMP-2
hybridization signal was then normalized to the $-actin hybridization signal to
account for any uneven loading of the gels. The expression of MMP-2 was 2.5 to
4.9 times higher than that of B-actin in the late secretory phase (Figure 7.8 panel c)
(Irwin 1996). The increased expression of MMP-2 in these late stages further
supports the role of MMP in endometrial cycling.

A number of other reproductive events have been associated with the
expression of MMPs and their inhibitors. Ovulation and the release of a mature
ovum require the coordination of MMPs and their inhibitors (TIMPs). Implantation
appears to involve the specific expression of MMP-9 by the embryo and the
modulation of invasion with TIMP production by the decidua (Graham 1994). These
reproductive events have the common theme of ECM turnover, and it is suspected
that modulation of MMPs play an important role in the body’s control over these

reproductive functions.

Regulation of mammary involution

In normal mammary glands, stromelysin-l (MMP-3) is synthesized by
fibroblast but not by epithelial cells. This expression has been associated with ECM
remodeling in the breast, such as, mammary gland ductal branching and mammary
gland involution upon the cessation of lactation. The expression of MMPs increases

during mammary involution (Talhouk 1992).

153

MMP-2
abcdefghijkl kb

. ..... - .. —2°9
1 . I .3 V
' MCcarlLlatunidJmJat: Dec Pia
proliferative secretory
B-actin

abcdefghijkl kb

 

12.3 -: ; 3..
. MC .
proliferative secretory

Densitometry
abcdefghijkl

 

.5
fl
0
E
‘1‘
a.
2
2
9mm mew Dec Pla
proliferative secretory

Figure 7.8. Expression of MMP-2 mRNA in human endometrium throughout the menstrual
cycle. Samples of total RNA(20 ug) extracted from the following tissues were analyzed by
Northern blot hybridization: early, mid, and late proliferative endometrium (lanes a-d); mid
cycle endometrium (MC, lane e); early, mid, and late secretory endometrium (lanes f-j); early
pregnancy decidua (Dec, lane k); term placenta (Pla, lane 1). (A) Autoradiograph of
hybridization with the MMP-2 cDNA probe. (B) Autoradiograph of the same membrane
hybridized with the -actin cDNA probe. (C) Densitometric analysis of the autoradiographs
shown in A and B, using the integrated area under the respective absorbance curves to

calculate the MMP-2/-actin hybridization ratio for each sample (Irwin 1996).

154

During involution, restructuring of the mammary gland occurs as a
coordinated process of alveolar apoptosis and lobular-alveolar remodeling. The
removal of the suckling triggers this process. With the loss of suckling, milk
accumulates within the alveolar lumens, and the levels of systemic lactogenic
hormones falls (Feng 1995). Mammary gland involution goes through two distinct
phases. In the first phase alveolar cells undergo apoptosis, and there is no remodeling
of the lobular-alveolar structure. Also occurring in this phase is expression of
TIMP-1 and cell cycle control proteins (c-Jun, JunB, JunD, c-fos, and c-Myc) and
the decreased expression of milk protein genes. In the second phase, the
lobular-alveolar structure of the gland is obliterated as proteases degrade basement
membrane and ECM components. In phase two, gelatinase A (MMP-2),
stromelysin-l (MMP-3), and stromelysin-2 (MMP-10) and urokinase type
plasminogen activator are expressed (Li 1997). The progressive gain of death signals
and the loss of surviving factors control the two stages of mammary gland involution.
In the first stage, the build up of milk within the alveolar cell triggers apoptosis but
also upregulates the expression of survival factors that do not overcome apoptosis
but do protect the remodeling of the mammary gland (phase 2). These survival
factors persist for 48 hours, and are important for the restoration of lactation in
case suckling is renewed. After 48 hours the second phase begins and the activities
of MMPs and uPA result in the destruction of the basement membrane and involution
of the mammary gland. These studies suggest that a spatial and temporal
coordination in the expression of MMPs and TIMPs is important for mammary

involution.

155

Wound healing

Skin wound healing is a complex process characterized by re-epithelization
and restoration of underlying connective tissue. During this process, keratinocytes,
endothelial cells, and fibroblast proliferate and these cells as well as inflammatory
cells migrate to the site of injury where they interact with each other and with the
ECM (Okada 1997).

The migration of cells and the remodeling of tissues during wound healing
require the controlled turnover (degradation) of the ECM and the release of growth
factors. The expression of gelatinase A (MMP-2) has been localized in connective
tissue of healing wounds and the expression of gelatinase B (MMP-9) is associated
with migrating epithelial sheets during human skin wound healing (Sato 1994).

The expression of MMP RNA during rat skin wound healing has recently
been examined (Okada 1997). Several MMPs were discovered in healing wounds
by Northern blot analysis, such as, MT-MMP, gelatinase A (MMP-2), gelatinase B
(MMP-3), and collagenase 3 (MMP-8), stromelysin-1 (MMP-3) and stromelysin-3
(MMP-ll) (Figure 7.9 panel A). The expression of inhibitors such as TIMP-l,
TIMP-2 and TIMP-3 were also measure by Northern blot analysis (Okada 1997).
The expression patterns of these enzymes and their inhibitors provide a number of
very interesting clues to the role of MMPs in healing wounds. The coordinate
expression of gelatinase A, MTl-MMP, and TIMP-2 suggest that the trimolecular
activation of MMP-2 at the surface of cells is an important mechanism in wound
healing. Studies have suggested that stromelysin-l is the activator of progelatinase
B. Data given in Figure 7 .9 (panel A) suggest that this mechanism of activation
may be relevant in wound healing. These findings were supported by transfection
of the African green monkey kidney cell line, COS-l, with plasmids that encode for
MMPs that are expressed in the healing wounds of rats (MTl-MMP, MMP-2,
MMP-9, MMP-3, MMP-11, and MMP-8) (Okada 1997). COS-1 cells were also

156

IIT1-IIHP

601A

573

ST1

6018

0013

3684

 

3 oars AFTER wounmuo
C

2 1

3 5 7 1014

~nn-

”IIIIH
“Pu-o

'0

HUI
13

«Ill!-

1234567

kb

‘- 4.2

‘- 3.1

‘- 2.4

«1.8

«2.6

«1.5

"MP1

TIMP2

TIMP3

3684

DAYS AFTER WOUNDING

 

NORMAL

13571014 kb

 

1234567

Figure 7.9. Northern blot analysis of MMP and TIMP RNAs during rat skin wound healing.
Total RNA (10 ug) from normal skin (lanes 1) and skin wounds on days 1, 3, 5, 7, 10, and 14
afier cutaneous incision (lanes 2-7), were electrophoresed, transferred to nylon membranes,
and hybridized with 32P-labeled cDNA probes for rat MTl-MMP, GelA, ST3, STl , GelB,
Col3 (A); TIMPl , TIMP2, TIMP3 (B). Blots were reprobed with the 36B4 cDNA used as a
loading control (Okada 1997).

157

transfected with TIMP-l, TIMP-2, or TIMP3 cDNA sequence (Okada 1997).
Conditioned medium samples were collected from these transfected cells and
analyzed by zymography (Okada 1997). These data indicate that gelatinase A is
most efficiently activated by MT] -MMP and most efficiently inhibited by TIMP-3.
The most efficient activator of gelatinase B was Stromelysin-1, and all TIMP
appeared to be successful at inhibiting the activity of MMP-9 (gelatinase B) (Okada
1997).

The regulation of MMP expression in wound healing requires tight regulation.
Any loss in the balance of the MMP to TIMP ratio can lead to uncontrolled
degradation of ECM in chronic wounds. There is an association of increasing age
with chronic wound disorders. One study recently examined the expression patterns
of collagenase-l, gelatinase A, gelatinase B, and Stromelysin-l in 132 healthy
humans between the ages of 19 and 96 years (Ashcroft 1997). These individuals
underwent 4-mm thick skin punch biopsies followed by wound excision between
day 1 and day 180. The results from this study indicated that the expression of
MMP-2 and MMP-9 were increased in older patients. Therefore, age related
alteration in the control of MMP expression may be important in the structural
changes observed with age in the skin and in rendering old skin susceptible to a

variety of pathologies resulting in impaired and chronic wound healing.

Nerve regeneration

Skin and brain have a great deal in common because of their ectodermal
origin. Several common mechanisms seem to underlie their normal structuring during
development and remodeling following trauma. Several assumptions are required
when considering mechanisms of actions in nerve regeneration (Brodkey 1993).
Most of the molecules involved in nerve regeneration are the same molecules seen

in similar regenerative processes throughout the body, namely MMPs, TIMPs,

158

cytokines, and growth factors. These molecules are under developmental control
in that they are expressed during critical periods of development, and normally
downregulated in the adult. These molecules often reappear following injury.
Finally, the signals that lead to the expression or reappearance of MMPs, TIMPs,
cytokines, and growth factors in nerve regeneration may be similar to the rest of
the body.

A remodeling process begins after injury to a peripheral nerve. This process
leads to the degeneration and regeneration of the axon (La Fleur 1996). After the
injury, axons in the distal segment undergo degeneration, which involves the removal
of axonal and myelin debris. Phagocytic cells remove the degenerating axons and
myelin, while dividing Schwann cells remain within the basement membrane tube
that surrounded the original nerve fiber. When regenerating axons from the proximal
segment re-ente'r the peripheral nerve matrix, Schwann cells ensheath and
re-myelinate them. The regenerating axons proceed to grow within the intact
Schwann cell-derived tubes.

The recruitment of macrophages appears two to three days after the injury.
Macrophages express several MMPs including gelatinase B (MMP-9) and
Stromelysin-l (MMP-3). These MMPs aid in the removal of axonal and myelin debris.
These phagocytes also produce TGF-B, TNF-a and IL-1, which induce Schwann
cells and fibroblasts to secrete nerve growth factor B (N GF-B) and TIMP-l. These
factors increase the rate of regeneration (La Fleur 1996).

During the response to nerve injury, infiltrating macrophages secrete
gelatinase B and resident Schwann cells are stimulated to produce TIMP-l. The
production of gelatinase B and TIMP-l in close proximity implies that ECM
degradation is restricted to small-defined sites near the cell surface. Gelatinase B
(MMP-9) can degrade myelin basic proteins. Increased proteolytic activity may

enhance meylin degradation during degenerative stages, or release Schwann cells

159

from their basement membrane connections during the proliferation stages, and lead
to the reestablishment of axons. The expression of TIMP-1 ensures the preservation
of the basement membrane, despite the presence of high levels of proteases released
after nerve injury. TIMP-1 protects the Shwann cell basement membrane from
uncontrolled degradation by gelatinase B (MMP-9) and stromelysin-l (MMP-3).
The basement membrane plays an important role in the maintenance of tissue
structure and orderly reconstruction following injury by serving as a scaffold for
the cellular migration, arrangement or attachment. Thus, the basement membrane
is critical for the promotion and guidance of axonal regrowth. Regenerating axons
grow along the inner surface of the Schwann cell basement membrane. The basement
membrane also provides the columnar organization of multiplying Schwann cells
during the repair process. In addition to the structural support, the basement

membrane also provides a favorable substrate for axonal growth.

PATHOLOGICAL CONDITIONS

There are several pathological conditions that involve the turnover of ECM.
In many of these conditions, MMPs have been implicated as major contributors to
ECM turnover by virtue of their ability to degrade components of the ECM and
their expression in diseases such as, arthritis cancer and glomerular nephropathy.

This section highlights some interesting pathologies associated with MMPs.

Arthritis

A great deal of research has gone into determining the role of MMPs in the
complex cascade of proteolytic events which results in tissue destruction involved
in arthritis. There is strong evidence to suggest that cartilage destruction is the
result of an imbalance in the MMP to TIMP ratio. The production of MMP and

TIMP in cultured synovial explants from rabbits with experimentally induced arthritis

160

 

showed that induction of the lesion correlated with a reduction in TIMP and an
increase in MMP level (Murphy 1991). Treatment with corticosteroids inhibited
the development of arthritis and suppressed the synthesis of MMP in a rabbit model
of arthritis (Murphy 1991). Elevated levels of MMP have been extracted from both
human osteoarthritic cartilage and the cartilage of dogs with experimentally induced
articular changes resembling human arthritis (Martel-Pelletier 1988). Increased
MMP expression has also been associated with the appearance and progression of
the degenerative changes found in aging cartilage (Martel-Pelletier 1988).
Rheumatoid arthritis is a chronic polyarthritis of unknown etiology affecting
1% of the World’s population. Features characteristic of rheumatoid arthritic joint
destruction, including bone absorption and synovial overgrowth, can be
experimentally reproduced by augmentation of c-fos gene expression. A study
concerning the role of the c-fos proto-oncogene in rheumatoid arthritis was recently
preformed in mice (Shiozawa, 1997). This study was of particular interest because
of the importance of c-fos/c-jun heterodimer (AP-1) in the regulation of the
expression of IL-lB, IL-6, TNF-a and most MMPs. To test the requirement of
c-fos gene expression in arthritic joint destruction, ongoing arthritis was inhibited
by blocking the AP-l signal. Short double stranded DNA oligonucleotides containing
AP-l sequences (AP-l oligos) and the control DNA oligonucleotides of the same
length discordant only at the AP-l consensus sequences (control oligos) were
administered intraperitoneally to mice twice weekly from 14 to 42 days after the
first injection until examination of the joints. AP-l oligos compete for the binding
of AP-l in viva at the promoter consensus binding sites and block the induction of
the MMP gene by c-fos. Arthritic joint destruction was inhibited in a dose dependent

manner in mice treated with AP-l oligos (Shiozawa 1997). The expression of mRN As

161

for IL-lB, lL-6, TNF-or, stromelysin l (MMP-3), and the 92-kDa gelatinase
(MMP-9) were reduced in mice receiving AP-l oligos compared to control oligos
(Figure 7.10) (Shiozawa 1997). This model helps to identify the pathogenesis of

rheumatoid joint destruction. .

Cancer

Matrix metalloproteinases are thought to play an important role in cancer
invasion and metastasis since they have the ability to degrade the basement membrane
and other components of the ECM (Webber 1995). Degradation of the basement
membrane and invasion into the stroma is likely mediated by MMPs and is observed
in cancer of the colon, prostate, breast and uterus (Crawford 1995, Lochter 1997,
Sterns 1993, Webber 1995). The ECM degradation by MMPs is controlled at the
same basic levels previously review in this chapter, expression, activation, and
inhibition.

Expression of MMPs in cancer cells can be induced by a variety of growth
factors, cytokines, oncogenes, and tumor promoters (Martrisian 1992). MMPs are
responsive to growth factors such as EGF, PDGF, bFGF, and TGF-a, which are
produced by numerous tumors in vivo. TNF-a and IL-1 B are produced by infiltrating
eosinophils and macrophages and have also been show to induce MMPs in fibroblast
cells (Chambers 1995). One recent study in human giant cell tumors (GCT) of the
bone found evidence suggesting that these cells secrete cytokines that upregulate
the expression of gelatinase A by stromal cells (Rao 1997).

Steroid Hormones appear to down regulate MMP expression in prostate
(Schneikert 1996). Several gelatinolytic proteinases have been detected by
zymographic analysis of extract of rat ventral prostate. The gelatinolytic activity
increased after castration and decreased upon treatment with testosterone

(Schneikert 1996). Studies using human prostatatic cell lines, DU145 and LNCaP,

162

 

 

lL-l B TNFa 1L-6 MMP-2 MMP-3 MMP-9

I I r I rj

Control DNA oligos #1 95.32;;
# 2

 

Figure 7.10. Quantification of mRN A for cytokine and matrix metalloproteinase in synovia
of mice treated with AP-l DNA oligos or control DNA oligos. mRNA of mice is reversibly
transcribed and coamplified relevant genes are quantified with reference to coamplified G3PDH
mRN A using Amplisensor assay. The amount of mRN A is expressed with an arbitrary scale
for individual cytokines and MMPs using serial dilution of one representative sample as
standard. The differences betwee two groups, except for MMP-2 and IL-1 , are < 0.02, and
those of IL-1 are < 0.06 by the Student’s nonpaired T test (Shiozawa 1997).

163

show that androgens negatively regulate the expression of interstitial collagenases
(MMP-1), stromelysin l (MMP-3), and matrilysin (MMP-7) (Schneihert 1996).
Androgen receptors appear to exert this effect on the ets family of transcription
factors. In contrast, the glucocorticoid receptor has no interaction with ets
transcription factors, and exerts its negative regulation by interaction with the
transcription factor AP-l, an important regulator of several MMP genes.

The expression of MMP-2 was examined in patients with benign prostatic
hyperplasia (BPH) and varying Gleason grades of malignant prostate cancer (Stearns
1993). The Gleason grade is a tool for rating the aggressiveness of malignant
prostate tumor, where scores are given 1-10 with 10 being the most aggressive.
These studies showed weak or no staining of gelatinase A in patients with BPH.
Patients with malignant prostatic cancer show increasing expression of gelatinase
A as Gleason scores increased. These data suggest MMP-2 expression is indicative
of an invasive phenotype (Stearns 1993).

The balance of MMP to TIMP is also important in cancer. Upregulation of
TIMP-1, by cytokines has been observed (Stearns 1995). The PC-3 human prostatic
carcinoma cell line showed low basal expression of TIMP-1. Treatment with varying
doses of IL-6 and IL-10 resulted in a dose dependent increase in TIMP expression.
Conversely, IL-1 and TNF-a decrease TIMP-1 expression while inducing MMP-2
in fibroblast (Stearns 1995). These results suggest that the balance of TIMP-1 and
MMP-2 is important in the progression of prostate cancer and that this balance is
somehow mediated by paracrine or autocrine cytokine stimuli.

MT-MMPs can activate progelatinase A in breast adenocarcinomas. In vivo
observations demonstrate that stromal cells are the main source of MMP expression.
More precisely, in breast cancer, MT-MMP and MMP-2 are localized in stromal
fibroblast next to the invading fronts of tumor cells. This specific location of

expression suggests that tumor cell may induce stromal cells to overexpress

164

MT-MMPs by various difussable growth factors. An investigation, using two
different breast cancer cell lines, one non-invasive (MCF 7) and the other invasive
(MDA-MB-231), was conducted to compare their ability to induce MT-MMP
production in human fibroblasts (Polette 1997). Conditioned medium from MD
A-MB-231 was able to induce the fibroblast to express MT-MMP and activate
MMP-2. In contrast, MCF-7 condition medium had no effect on the fibroblasts.
Gelatinase A and MT-MMP have been associated with invadopodia. As previously
described, MT-MMP are highly implicated in the activation of MMP-2 and in situ
hybridization studies have associated MMP-2 and MT-MMP to the invadopodia of

breast cancer cells (Thompson 1994).

Glomerular Nephropathy

The balance of MMP to TIMP is critical to the function of the kidney
glomerulus. The glomerular filtration is dependent on the structure provided by
the ECM. Abnormal expansion of the ECM can lead to chronic sclerotic states,
while increased degradation of ECM can result in acute nephritis. The glomerular
mesangial cell (GMC) is centrally involved in both chronic sclerotic and acute
inflammatory processes characteristic of many forms of renal disease. Glomerular
mesangial cell activation involves augmented proliferation and synthesis of ECM
proteins, specifically type IV collagen, laminin and hepran sulfate (Yasuda 1996).
Following the activation of GMCs there is an enhanced secretion of gelatinase A by
these cells (Turck 1995). This close temporal linkage between GMC activation and
GMC secretion of gelatinase A suggests that the turnover of ECM is dependent on
the balance between its synthesis and its degradation. Consequently, glomerular

sclerosis may be due to increased synthesis of ECM proteins and/or reduced protease

165

activity, by either decreased expression or increased inhibition. In contrast,
glomerular nephritis can be induce by decreased ECM protein synthesis and/or

increased protease activity by either increase expression or decreased inhibition.

One recent study examined cultured mesangial cells that overexpressed
gelatinase A (Turck 1996). These cells accurately reflected an inflammatory
phenotype: they had increased gelatinase A expression, their morphology included
lamellopodia, and they did not synthesize IV collagen (Turck 1996). This
inflammatory phenotype could be restored to a quiescent phenotype with the
suppression of gelatinase A expression by transfection with episomal antisense
gelatinase A gene or hammerhead ribosome targeted to gelatinase A mRNA. The
genetically altered cells reflected a quiescent phenotype: they did not express
gelatinase A, their morphology was rounded, and they synthesized type IV collagen
(Turck 1996).

Passive Heymann’s nephritis is a glomerular nephritis that can be induced in
rats by visceral epithelial injury with injections of 10 : g of a sheep antibody,
anti-FxlA (McMillian 1996). An investigation examined the expression of gelatinase
B in normal rats and those injected with experimentally induced Heymann’s nephritis
at five and fifteen day after injury (injection with sheep antibody) (McMillian 1996).
Immunohistochemical staining for gelatinase B antigen was performed on glomerular
sections with normal cells and passive Heymann’s nephritis cells five and fifteen
days after injection of sheep antibody. Normal cells showed very low levels of
gelatinase B antigen either within the glomeruli tubules or interstitium. In contrast,
passive Heymann’s nephritis cells showed a progressively increasing expression
pattern. At day five, gelatinase B antigen was primarily expressed in the visceral
epithelium. At day fifteen, gelatinase B antigen reactive signal was stronger in the

visceral epithelium and was also seen in some of the endothelial in the capillary

166

 

lumens (McMillian 1996). These data indicate that gelatinase B is not only expressed
by stromal cell, but in the case of nephritis, gelatinase B can also be expressed by
visceral epithelium.

In contrast to the increased expression of MMPs, as seen in glomerular
nephritis, glomerular sclerosis appears to involve a decrease in MMP activity.
Glomerular sclerosis ensues with the accumulation of glomerular ECM. It is possible
that an imbalance of the MMP to TIMP ratio favoring TIMP might be involved in
this accumulation of ECM. To examine this possible decrease in MMP activity and
or increase in TIMP activity, studies have used the Obese Zucker rat model (Schaefer
1997).

The obese Zucker rat is a model of obesity and hyperlipidemia with many
similarities to human non-insulin dependent diabetes melitus. This metabolic disease
results in the development of albuminuria and mesangial matrix expansion in the
glomerulus. The accumulation of matrix components like type IV collagen,
fibronectin, laminin and proteoglycans results in glomerulosclerosis. These rats
exhibit a decrease in glomerular gelatinolytic activity due to reduced gelatinase B
expression and increased expression of TIMP-1. The glomerular gelatinase B mRNA
is reduced 5 times, while the expression of TIMP-1 , the natural inhibitor of gelatinase
B, is enhanced (Scharfe 1997). This rat model shows that glomerular sclerosis can
result from the build up of ECM which is meditated by both a loss of gelatinase B

expression and increase in the expression of TIMP-1.

APPLICATIONS

Recent epidemiological studies have shown a 40 to 50% reduction in mortality
from colorectal cancer in individuals who take non-steroidal anti-inflammatory drugs
(NSAIDs) regularly. Most NSAIDs inhibit both cyclooxygenase COX-1 and

COX-2. COX-l is expressed in normal intestine, however, COX-2 is undetectable

167

 

in normal intestine, but is elevated in 85% of colorectal adenocarcinomas. Human
colon cancer cells, Caco-2, were tranfected to express COX-2 (Tsujii 1997). These
cells showed an increased expression of MT-MMP and increased activation of
MMP-2 (gelatinase A). The controls were transfected with vector only, and did
not show an increase in the expression of MT-MMP or increased in MMP-2 activity.
The COX-2 expressing Caco-2 cells showed increased invasion. This invasive
activity could be reversed by sulindac sulfide, a known COX inhibitor. The
zymogram in Figure 7.11 shows the reversible activation of MMP-2 (Tsujii 1997).
These data suggest that the COX-2 induces the production of MT-MMP which
activates MMP-2 and result in increased invasion. This action was reversed by
treatment by a NSAID.

Human glioblastoma cell lines express gelatinase A and gelatinase B at much
higher levels than non-cancerous gliomas and normal brain (Chintala 1997). After
treatment with cis-diamminedichloroplatinum (cisplatin), glioblastomas cells were
examined by gelatin zymography. Gelatinase expression was decreased, however
the mechanism for this decrease is not yet known. This study indicates that treatment
with cisplatin might reduce invasion by decreasing gelatinase activity, in addition
to it cytotoxic effects (Chintala 1997).

Studies examining human tumors and treatment with synthetic inhibitors of
matrix metalloproteinases have shown promise (Santos 1997). Batimastat (BB94),
one such inhibitor, has been useful in treating patients with breast cancer. Marimastat
(BB-2516), another MMP inhibitor, has resulted in reductions of circulating breast
tumor antigens, an indication of anti-tumor efficacy. Due to the lack of cytotoxicity
of MMP inhibitors and their unique mode of action, this new class of compounds
may provide new treatment for cancer patients. New synthetic MMP inhibitors are
undergoing laboratory investigation (Santos 1997, An 1997). The MMP synthetic

inhibitor, CT1746, has been showed to be effective against gelatinase A, gelatinase

168

Inactive MMP-2 —> ~~ .,. .._.. Gelatin
Active MMP-2 —-> “" Zymography

MW. MMP-2 _.., ”a“ W9516m
Active MMP-2 -—- — ~ Blot

Northern

MT-MMP-t -> w ‘3“ Blot

 

Figure 7.11. Matrix MMP-2 and MT—MMP expression in Caco-2 cells. (T0p) Gelatin
zymography of proteins secreted into the cell culture medium. Lane 3 demonstrates the
result following treatment with sulindac sulfide (SS; 25 micro M) for 24 hr. A 68- and
62-kDa band are seen in the COX-2-expressing Caco-2 cells; whereas, only the inactive
68-kDa band is een in the control Caco-2 cells. (Middle) Immunoblotting results with an
anti-MMP-2 antibody in an identical experiment to that shown in the Top panel. This result
confirms the gelatin zymography data shown in cells which is inhibited in lane 3 by treatment
with 25 microM sulindac sulfide for 6 hr. (Tsujii 1997).

169

B, and stromelysins (An 1997). Oral administration of CT] 746 to nude mice, bearing
tumors from the Co-3 human colon cancer cell line, resulted in the reduction of
tumor size and a decrease in tumor spread. Treatment with this drug significantly
prolonged median survival time in tumor bearing mice from 51 to 78 days (An
1997). These data suggest that synthetic inhibitor of MMP might be useful as

non-cytotoxic chemopreventive or chemotherapeutic agents.

CONCLUSIONS

Matrix metalloproteinases are a family of enzymes that have evolved to digest
specific extracellular matrix components. The activity of these proteases is regulated
at three levels, expression, activation, and inhibition. These points strongly support
the role of MMPs and TIMPs as important modulators of ECM turnover. Several
condition both normal and pathologic involve this ECM turnover. The evidence
discussed in this chapter demonstrates common mechanisms of matrix turnover.
These include matrix metalloproteinases, their inhibitors and their ECM substrates,
in normal development, wound healing and regeneration, as well as pathological

conditions, such as, arthritis, glomerular nephropathy and cancer.

170

REFERENCES

An, Z., Wang, X., \Vrllmott, N., Chander, S. K., Tickle, S., Docherty, A. J ., Mountain, A.,
Millican, A. T., Morphy, R., Porter, J. R., Epemolu, R. 0., Kubota, T., Moossa, A.
R., and Hoffman, R. M.: Conversion of highly malignant colon cancer from an
aggressive to a controlled disease by oral administration of a metalloproteinase
inhibitor. Clin Exp Metastasis 15: 184-195, 1997.

Apte, S. S., Fukai, N., Beier, D. R., and Olsen, B. R.: The matrix metalloproteinase-l4
(MMP-l4) gene is structurally distinct from other MMP genes and is co-expressed
with the TIMP-2 gene during mouse embryogenesis. J Biol Chem 272: 25511-25517,
1997.

Ashcroft, G S., Horan, M. A., Herrick, S. E., Tarnuzzer, R. W., Schultz, G S., and Ferguson,
M. W.: Age-related differences in the temporal and spatial regulation of matrix

metalloproteinases (MMPs) in normal skin and acute cutaneous wounds of healthy
humans. Cell Tissue Res 290: 581-591, 1997.

Basset, P., Wolf, C., and Chambon, P.: Expression of the stromelysin-3 gene in fibroblastic

cells of invasive carcinomas of the breast and other human tissues: a review. Breast
Cancer Res Treat 24: 185-193, 1993.

Brodkey, J. A., Gates, M. A., Laywell, E. D., and Steindler, D. A.: The complex nature of
interactive neuroregeneration-related molecules. Exp Neurol 123: 251-270, 1993.

Chambers, A. F. and Matrisian, L. M.: Changing views of the role of matrix metalloproteinases
in metastasis. J Natl Cancer Inst 89: 1260-1270, 1997.

Chintala, S. K., Ali Osman, F., Mohanam, S., Rayford, A., Go, Y., Gokaslan, Z. L., Gagercas,
E., Venkaiah, B., Sawaya, R., Nicolson, G L., and Rao, J. S.: Effect of cisplatin and
BCNU on MMP-2 levels in human glioblastoma cell lines in vitro. Clin Exp Metastasis .
15: 361-367, 1997.

Crawford, H. C. and Matrisian, L. M.: Tumor and stromal expression of matrix
metalloproteinases and their role in tumor progression. Invasion Metastasis 14:
234-245, 1995.

Goldberg, G I., Strongin, A., Collier, I. E., Genrich, L. T., and Marmer, B. L.: Interaction of
92-kDa type IV collagenase with the tissue inhibitor of metalloproteinases prevents

dimerization, complex formation with interstitial collagenase, and activation of the
proenzyme with stromelysin. J Biol Chem 267: 4583-4591, 1992.

171

 

Graham, C. H., Connelly, I., MacDougall, J. R., Kerbel, R. S., Stetler Stevenson, W. G, and
Lala, P. K.: Resistance of malignant trophoblast cells to both the anti-proliferative
and anti-invasive effects of transforming growth factor-beta. Exp Cell Res 214: 93 -99,
1 994.

Gross, J. and Lapiere, C. M.: Collagenolytic activity in amphibian tissue: a new tissue culture
assay. Proc Natl Acad Sci USA 48: 1014-1022, 1962.

Imai, K., Ohuchi, E., Aoki, T., Nomura, H., Fujii, Y., Sato, H., Seiki, M., and Okada, Y.:
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,
1 996.

Irwin, J. C., Kirk, D., Gwatkin, R. B., Navre, M., Cannon, P., and Giudice, L. C.: Human
endometrial matrix metalloproteinase-2, a putative menstrual proteinase. Hormonal
regulation in cultured stromal cells and messenger RNA expression during the
menstrual cycle [see comments]. J Clin Invest 97: 438-447, 1996.

Ishizuya-Oka, Atsuko, Ueda, Shuichi, and Shi, Yun-Bo. Transient expression of stromelysin-3
mRNA in the amphibian small intestine during metamorphosis Cell Tissue Research
283: 325-329 1996.

La Fleur, M., Underwood, J. L., Rappolee, D. A., and Werb, Z.: Basement membrane and
repair of injury to peripheral nerve: defining a potential role for macrophages, matrix
metalloproteinases, and tissue inhibitor of metalloproteinases-l. J Exp Med 184:
2311-2326, 1996.

Lee, A. Y., Akers, K. T., Collier, M., Li, L., Eisen, A. Z., and Seltzer, J. L.: Intracellular
activation of gelatinase A (72-kDa type IV collagenase) by normal fibroblasts. Proc
Natl Acad Sci USA 94: 4424-4429, 1997.

Li, M., Liu, X., Robinson, G, Bar Peled, U., Wagner, K. U., Young, W. S., Hennighausen,
L., and Furth, P. A.: Mammary-derived signals activate programmed cell death during
the first stage of mammary gland involution. Proc Natl Acad Sci USA 94: 3425-3430,
1997.

Lochter, A., Galosy, S., Muschler, J ., Freedman, N., Werb, Z., and Bissell, M. J .: Matrix
metalloproteinase stromelysin-l triggers a cascade of molecular alterations that leads
to stable epithelial-to-mesenchymal conversion and a premalignant phenotype in
mammary epithelial cells. J Cell Biol 139: 1861-1872, 1997.

Matrisian, L. M.: The matrix-degrading metalloproteinases. Bioessays 14:455-463, 1992.

McMillan, J. 1., Riordan, J. W., Couser, W. G, Pollock, A. S., and Lovett, D. H.:
Characterization of a glomerular epithelial cell metalloproteinase as matrix

metalloproteinase-9 with enhanced expression in a model of membranous nephropathy.
J Clin Invest 97: 1094-1101, 1996.

172

Murphy, A. N., Unsworth, E. J ., and Stetler Stevenson, W. G: Tissue inhibitor of
metalloproteinases-2 inhibits bFGF-induced human microvascular endothelial cell
proliferation. J Cell Physiol 1 57: 351-358, 1993 .

Noel, A. C., Lefebvre, 0., Maquoi, E., VanHoorde, L., Chenard, M. P., Mareel, M., Foidart,
J. M., Basset, P., and Rio, M. C.: Stromelysin-3 expression promotes tumor take in
nude mice. J Clin Invest 97: 1924-30, 1996.

Ohuchi, E., Imai, K., Fujii, Y., Sato, H., Seiki, M., and Okada, Y.: Membrane type 1 matrix
metalloproteinase digests interstitial collagens and other extracellular matrix
macromolecules. J Biol Chem 272: 2446-2451, 1997.

Okada, A., Tomasetto, C., Lutz, Y., Bellocq, J. P., Rio, M. C., and Basset, P.: Expression of
matrix metalloproteinases during rat skin wound healing: evidence that membrane
type-1 matrix metalloproteinase is a stromal activator of pro-gelatinase A. J Cell Biol
137: 67-77, 1997.

Polette, M., Gilles, C., Marchand, V., Seiki, M., Tournier, J. M., and Birembaut, P.: Induction
of membrane-type matrix metalloproteinase 1 (MTl-MMP) expression in human
fibroblasts by breast adenocarcinoma cells. Clin Exp Metastasis 15: 157-163, 1997.

Rao, V. H., Singh, R. K., Bridge, I. A., Nefi‘, J. R., Schaefer, G B., Delimont, D. C., Dunn,
C. M., Sanger, W. G, Buehler, B. A., Sawaya, R., Nicolson, G L., and Rao, J. S.:
Regulation of MMP-9 (92 kDa type IV collagenase/gelatinase B) expression in stromal
cells of human giant cell tumor of bone. Clin Exp Metastasis 15: 400-409, 1997.

Santos, 0., McDermott, C. D., Daniels, R. G., and Appelt, K.: Rodent pharmacokinetic and
anti-tumor efficacy studies with a series of synthetic inhibitors of matrix
metalloproteinases. Clin Exp Metastasis 15: 499-508, 1997.

Schaefer, L., Han, X., August, C., Matzkies, F., Lorenz, T., and Schaefer, R. M. Differential
regulation of glomerular gelatinase B (MMP-9) and tissue inhibitor of
metalloproteinase-l (TIMP-l) in obses Zucker rats. Diabetologia 40: 1035-1043,
1997.

Schneikert, J ., Peterziel, H., Defossez, P. A., Klocker, H., Launoit, Y., and Cato, A. C.:
Androgen receptor-Ets protein interaction is a novel mechanism for steroid

hormone-mediated down-modulation of matrix metalloproteinase expression. J Biol
Chem 271: 23907-23913, 1996.

Schorpp, M., Mattei, M. G, Herr, I., Gack, S., Schaper, J., and Angel, P.: Structural
organization and chromosomal localization of the mouse collagenase type I gene.
Biochem J 308: 211-217, 1995.

173

 

Shiozawa, S., Shimizu, K., Tanaka, K., and Hino, K.: Studies on the contribution of c-fos/
AP-l to arthritic joint destruction. J Clin Invest 99: 1210-1216, 1997.

Stearns, M. E. and Wang, M.: Type IV collagenase (M(r) 72,000) expression in human
prostate: benign and malignant tissue. Cancer Res 53: 878-883, 1993.

Stearns, M., Wang, M., and Stearns, M. E.: Cytokine (IL-10, IL-6) induction of tissue inhibitor
of metalloproteinase 1 in primary human prostate tumor cell lines. Oncol Res 7:
173-181,1995.

Takino, T., Sato, H., Shinagawa, A., and Seiki M.: Identification of the second membrane-type
matrix metalloproteinase (MT-MMP-Z) gene from a human placenta cDNA library.
MT-MMPs form a unique membrane-type subclass in the MMP family. J Biol Chem
270: 23013-23020, 1995.

Talhouk, R. S., Bissell, M. J ., and Werb, Z.: Coordinated expression of extracellular
matrix-degrading proteinases and their inhibitors regulates mammary epithelial function
during involution. J Cell Biol 118: 1271-1282, 1992.

Thompson, E. W., Yu, M., Bueno, J ., Jin, L., Maiti, S. N., Palao Marco, F. L., Pulyaeva, H.,
Tamborlane, J. W., Tirgari, R., and Wapnir, I..: Collagen induced MMP-2 activation
in human breast cancer. Breast Cancer Res Treat 31: 357-370, 1994.

Tsujii, M., Kawano, S., and DuBois, R. N.: Cyclooxygenase-Z expression in human colon
cancer cells increases metastatic potential. Proc Natl Acad Sci USA 94: 3336-3 340,
1997.

Turck, J ., Pollock, A. S., Lee, L. K., Marti, H. P., and Lovett, D. H.: Matrix metalloproteinase
2 (gelatinase A) regulates glomerular mesangial cell proliferation and differentiation.
J Biol Chem 271: 15074-15083, 1996.

Webber, M. M. and Waghray, A.: Urokinase-mediated extracellular matrix degradation by
human prostatic carcinoma cells and its inhibition by retinoic acid. Clinical Cancer
Research 1: 755-761 1995.

Yasuda, T., Kondo, S., Homma, T., and Harris, R. C.: Regulation of extracellular matrix by

mechanical stress in rat glomerular mesangial cells. J Clin Invest 98: 1991-2000,
1996.

174

 

CHAPTER 8

SIGNIFICANCE AND APPLICATION OF MNIPs AND TIMPs

SIGNIFICANCE

The previous chapters have discussed the matrix metalloproteinases, their
substrates, and their inhibitors. The functional role of matrix metalloproteinases
(MMPs) and tissue inhibitors of MMP (TIMPs) is to modulate extracellular matrix
(ECM) turnover in several normal and pathological processes. An intricate balance
between MMP and TIMP is seen in normal processes such as development and
wound healing. However, a disruption in this balance of MMP and TIMP is observed
in important human diseases such as arthritis and cancer. As we approach the new
millenium, the percentage of individuals affected by these diseases is increasing
(Powell 1993). The knowledge compiled in this thesis can be applied to future

research and treatment of these diseases.

ARTHRITIS

Much recent research has been directed towards determining the role of
MMPs in the arthritis. A central role for MMPs produced by endogenous tissue
cells as well as those of inflammatory cells is n0w evident. Experimental models
have shown that cartilage destruction is the result of an imbalance in favor of MMPs
over TIMPs. Increased expression of MMPs and decreased expression of TIMPs
has been observed in cultured human tissue cells from rheumatoid joints (Murphy
1991). These studies suggest that tissue destruction, associated with arthritis, is

due to an insufficient amount of TIMPs to counteract the destructive effects of

175

MMPs. Current corticosteroid treatment for arthritis not only blocks inflammation
but also appears to inhibit the expression of MMPs (Murphy 1995). Tetracycline
inhibits the activity of MMPs, and can reduce the severity of osteoarthritis in dogs
(Yu 1992). This inhibition of MMP by tetracycline is independent of its antimicrobial
activity and may be mediated by the chelation of zinc and calcium (Suomalainen
1992). New synthetic inhibitors of MMP are being developed as potential therapeutic
agents for arthritis (Docherty 1992, Murphy 1991, 1995).

CANCER

Metastasis or the formation of secondary tumors at distant sites from the
primary tumor is a major cause of cancer mortality. The failure of most current
therapies to successfully treat metastases stresses the urgency of understanding the
underlying mechanisms of cancer metastasis. There are several steps in the
progression of a primary tumor cell toward a metastatic phenotype (Chambers 1997).
Genetic manipulation of MMP and TIMP expression in several tumor cell lines has
demonstrated the involvement of these enzymes in tumor invasion and metastasis
(An 1997, Aoudjit 1998, Noel 1996). The imbalance of MMP to TIMP in favor of
MMP is a major factor that leads to metastasis (Matrisian 1992, Montironi 1996).
It has been suggested that retinoic acid can be useful in cancer prevention (Webber
1995, 1996). Considerable scientific evidence has demonstrated that retinoids are
effective in the prevention of a variety of cancers and in the inhibition of cancer
progression and invasion (Kim 1995, Waghary & Webber 1995). Treatment of
prostate cancer cells with retinoids decreases the net proteolytic activity of
urokinase-type plasminogen activators as well as MMPs (Waghray & Webber 1995,
Webber 1995, 1996). Urokinase is implicated in the activation of MMPs, therefore,
treatment with retinoids may decrease MMP activity. There is also recent evidence

to suggest that retinoids may modulate the expression of some MMP genes. These

176

studies indicate the need for further research on the effects of retinoids on cancer
metastasis. There are several synthetic inhibitors of MMPs currently under clinical
investigation, and include Batimastat and Marimastat (Santos 1997). These drugs
show excellent promise for the treatment of human breast cancers and pharmaceutical
companies are rushing to develop similar compounds for use as anticancer agents
(An 1997). Indeed, the study of MMPs and TIMPs will provide valuable tools by

which we can combat cancer.

CONCLUSIONS

Matrix metalloprotinase and TIMPs are important modulators of extracellular
matrix turnover. The process of ECM turnover is critical to a number of human
physiological conditions including normal processes such as development and wound
healing, and pathological processes, such as, arthritis and cancer. It is apparent
that the balance of MMP to TIMP is important for these processes, however the
regulation of MMP to TIMP balance is not yet fully understood. There are many
great challenges to be faced in the elucidation of the mechanisms that control the
balance of MMPs and TIMPs. Considering the application of this knowledge toward
the treatment of human disease, the rewards for meeting these challenges are equally

great.

177

REFERENCES

An, Z., Wang, X., Willmott, N., Chander, S. K., Tickle, S., Docherty, A. J ., Mountain, A.,
Millican, A. T., Morphy, R., Porter, J. R., Epemolu, R. 0., Kubota, T., Moossa, A.
R., and Hoffman, R. M.: Conversion of highly malignant colon cancer from an
aggressive to a controlled disease by oral administration of a metalloproteinase
inhibitor. Clin Exp Metastasis 15: 184-195, 1997.

Aoudjit, F., Potworowski, ER, and St-Pierre, Y.: The metastatic characteristics of murine
lymphoma cell lines in vivo are manifested after target organ invasion. Blood
91:623-629, 1998.

Chambers, A. F. and Matrisian, L. M.: Changing views of the role of matrix metalloproteinases
in metastasis. J Natl Cancer Inst 89: 1260-1270, 1997.

Crawford, H. C. and Matrisian, L. M.: Mechanisms controlling the transcription of matrix
metalloproteinase genes in normal and neoplastic cells. Enzyme Protein 49: 20-3 7,
l 996.

Crawford, H. C. and Matrisian, L. M.: Mechanisms controlling the transcription of matrix

metalloproteinase genes in normal and neoplastic cells. Enzyme Protein 49: 20-3 7,
1996.

Docherty, A. J., O’Connell, J., Crabbe, T., Angal, S., and Murphy, G: The matrix
metalloproteinases and their natural inhibitors: prospects for treating degenerative
tissue diseases. Trends Biotechnol 10: 200-207, 1992.

Guerin, E., Ludwig, M.G, Basset, P., and Anglard, P.: Stromelysin-3 induction and interstitial
collagenase repression by retinoic acid. Therapeutical implication of receptor-selective
retinoids dissociating transactivation and AP- 1 -mediated transrepression. J Biol Chem
272: 11088-11095, 1997.

Kim, J. H., Tanabe, T., Chodak, G W., and Rukstalis, D. B.: In vitro anti-invasive effects of
N-(4-hydroxyphenyl)-retinamide on human prostatic adenocarcinoma. Anticancer
Res 15: 1429-1434, 1995.

Matrisian, L. M.: The matrix-degrading metalloproteinases. Bioessays 14: 455-463, 1992.

Montironi, R., Lucarini, G., Castaldini, C., Galluzzi, C. M., Biagini, G., and Fabris, G.:
Immunohistochemical evaluation of type IV collagenase (72-kd metalloproteinase)
in prostatic intraepithelial neoplasia. Anticancer Res 16: 205 7-2062, 1996.

Murphy, G, Docherty, A. J ., Hembry, R. M., and Reynolds, J. J .: Metalloproteinases and
tissue damage. Br J Rheumatol 30 Suppl 1: 25-31, 1991.

178

Murphy, G: Matrix metalloproteinases and their inhibitors. Acta Orthop Scand Suppl 266:
55-60, 1995.

Noel, A.C., Lefebvre, 0., Maquoi, E., VanHoorde, L., Chenard, M.P., Marcel, M., Foidart,
J .M., Basset, P., and Rio, M.C.: Stromelysin-3 expression promotes tumor take in
nude mice. J Clin Invest 97: 1924-1930, 1996.

Powell, W. C., Knox, J. D., Navre, M., Grogan, T. M., Kittelson, J ., Nagle, R. B., and and
Bowden, G T.: Expression of the metalloproteinase matrilysin in DU-l45 cells

increases their invasive potential in severe combined immunodeficient mice. Cancer
Res 53: 417-422, 1993.

Santos, 0., McDermott, C.D., Daniels, R.G., and Appelt, K.: Rodent pharmacokinetic and
anti-tumor efficacy studies with a series of synthetic inhibitors of matrix
metalloproteinases. Clin Exp Metastasis 15: 499-508 1997

Suomalainen, K., Sorsa, T., Golub, L. M., Ramamurthy, N., Lee, H. M., Uitto, V. J ., Saari,
H., and Konttinen, Y. T.: Specificity of the anticollagenase action of tetracyclines:
relevance to their anti-inflammatory potential. Antimicrob Agents Chemother 36:
227-229, 1992 .

Waghray, A., and Webber, M. M.: Retinoic acid modulates extracellular urokinase-type
plasminogen activator activity in DU-145 human prostatic carcinoma cells. Clinical
Cancer Res 1: 747-753 1995.

Webber, M.M., and Waghray A.: Urokinase-mediated extracellular matrix degradation by
human prostate carcinoma cells and its inhibition by retinoic acid . Clin Cacner Res 1 :
755-761, 1995.

Webber, M. M., Waghray A., Bello, D., and Rhim, J. M.: Mini review: proteases and invasion
in human prostate epithelial cell lines; implications in prostate cancer prevention and
intervention. Radiation Oncol Invest 3: 358-362, 1996.

Yu, L.P. Jr, Smith, GN. Jr, Brandt,K.D., Myers, S.L., O’Connor, B.L., and Brandt, D.A.:
Reduction of the severity of canine osteoarthritis by prophylactic treatment with
oral doxycycline. Arthritis Rheum 35: 1150-1 159, 1992.

179