Matrilysin (matrix metalloproteinase 7) in parturition, premature

rupture of membranes, and intrauterine infection

Eli Maymon, MD,a Roberto Romero, MD,a, b Percy Pacora, MD,a Maria-Teresa Gervasi, MD,a
Samuel S. Edwin,a Ricardo Gomez, MD,a and David E. Seubert, MDb
Bethesda, Maryland, and Detroit, Michigan

OBJECTIVE: Matrix metalloproteinases are enzymes capable of degrading extracellular matrix components.
Matrilysin (matrix metalloproteinase 7), a novel member of this family, degrades fibronectin and proteoglycans.
The objective of this study was to determine whether parturition (either term or preterm), premature rupture of
the membranes, and microbial invasion of the amniotic cavity are associated with changes in the amniotic fluid
concentration of matrilysin.
STUDY DESIGN: A cross-sectional study was conducted with 275 women in the following categories:
(1) second trimester, (2) term not in labor, (3) term in labor, (4) term with microbial invasion of the amniotic cav-
ity, (5) preterm labor with intact membranes without microbial invasion of the amniotic cavity who delivered at
term, (6) preterm labor without microbial invasion of the amniotic cavity who delivered preterm, (7) preterm labor
with microbial invasion of the amniotic cavity, (8) preterm premature rupture of membranes with and without
microbial invasion of the amniotic cavity, and (9) term premature rupture of membranes not in labor and without
microbial invasion of the amniotic cavity. Matrilysin concentrations were measured with a sensitive specific
immunoassay that was validated for amniotic fluid.
RESULTS: Matrilysin was detectable in 97.4% (268/275) of the samples. The concentration of matrilysin in-
creased with advancing gestational age (r = 0.8; P < .001). Parturition at term was not associated with a signifi-
cant increase in amniotic fluid concentration of matrilysin. Preterm parturition in the absence of microbial inva-
sion of the amniotic cavity was associated with a significant increase in amniotic fluid concentration of matrilysin
(preterm labor with preterm delivery: median, 1.7 ng/mL; range, 0.45-21.6 mg/mL; vs preterm labor with term de-
livery: median, 1.2 ng/mL; range, 0.17-42.1 ng/mL; P < .05). Premature rupture of membranes without microbial
invasion of the amniotic cavity (either term or preterm) was not associated with a significant change in the amni-
otic fluid matrilysin concentration. Intra-amniotic infection was associated with a significant increase in amniotic
fluid matrilysin among both patients with preterm labor and patients with preterm premature rupture of mem-
branes (preterm labor with microbial invasion of the amniotic cavity: median, 3.2 ng/mL; range, 0.16-21.9 ng/mL;
vs preterm labor and delivery without microbial invasion of the amniotic cavity: median, 1.7 ng/mL; range,
0.45-21.6 ng/mL; vs preterm labor with term delivery: median, 1.2 ng/mL; range, 0.17-42.1 ng/mL; P < .01 for
each comparison; and preterm premature rupture of membranes without microbial invasion of the amniotic
cavity: median, 1.7 ng/mL; range, 0.29-13.9 ng/mL; vs preterm premature rupture of membranes with microbial
invasion of the amniotic cavity: median, 3.6 ng/mL; range, 0.59-20.3 ng/mL; P < .01).
CONCLUSION: Matrilysin is a physiologic constituent of amniotic fluid, and its concentration increases with ad-
vancing gestational age. Microbial invasion of the amniotic cavity in preterm gestations was associated with a
significant increase in amniotic fluid concentration of matrilysin. Matrilysin therefore may play a role in the host
defense mechanism. (Am J Obstet Gynecol 2000;182:1545-53.)

Key words: Intra-amniotic infection, matrilysin, matrix metalloproteinase, preterm labor

The uterine corpus, cervix, fetal membranes, and pla-
centa undergo extensive growth and remodeling during
From the Perinatology Research Branch, National Institute of Child
Health and Human Development,a and the Department of Obstetrics
and Gynecology, Wayne State University Hutzel Hospital.b
Presented at the Sixty-seventh Annual Meeting of The Central Associa-
tion of Obstetricians and Gynecologists, Maui, Hawaii, October 24-27,
1999.
Reprint requests: Eli Maymon, MD, Perinatology Research Branch, Na-
tional Institute of Child Health and Human Development, Wayne State
University Hutzel Hospital, Department of Obstetrics and Gynecology,
4707 St Antoine Blvd, Detroit, MI 48201.
6/6/107652
doi:10.1067/mob.2000.107652
pregnancy.1 Tissue restructuring requires both the syn-
thesis and degradation of the extracellular matrix. Matrix
metalloproteinases are a family of zinc-dependent en-
zymes capable of degrading such extracellular matrix
components as collagen, proteoglycans, and glycopro-
teins.2, 3 These enzymes have been implicated in remod-
eling during both normal and pathologic processes.4
Matrix metalloproteinase family members have typi-
cally been grouped according to substrate specificity.3
Matrilysin, also known as matrix metalloproteinase
(MMP) 7,4 was first discovered in the involuting uterus
and implicated in the breakdown of collagen during the
puerperium.5, 6 MMP-7 has been shown to degrade fi-

1545
1546 Maymon et al
bronectin, gelatin, laminin, elastin, casein, and proteo-
glycans, such as versican and aggrecan.1, 7-9 MMP-7 can
also activate the latent forms of the other matrix metallo-
proteinases and the proteolytic processing of other mole-
cules, such as tumor necrosis factor α, urokinase tissue
plasminogen activator, tissue-type plasminogen activator,
and β-cadherin.7, 10-12 Matrilysin (MMP-7) is a small (28-
kd) matrix metalloproteinase that has a unique structure:
it lacks the hemopexin-vitronectin binding domain char-
acteristic of the other matrix metalloproteinases.7 MMP-7
production is transcriptionally regulated, activated by en-
dopeptidases such as plasmin, and inhibited by tissue in-
hibitors of matrix metalloproteinases.
In the reproductive tract, high constitutive levels of ma-
trilysin and its messenger ribonucleic acid are found in
the epithelial cells of the uterus in rats and mice.6, 13, 14
Immunohistochemical and in situ hybridization studies
in humans have demonstrated the presence of protein
and messenger ribonucleic acid in the proliferative, late
secretory, and menstrual endometrium15, 16 and also in
syncytiotrophoblasts during early gestation and in cy-
totrophoblasts by the third trimester.16
The detection of fibronectin in vaginal and cervical se-
cretions is a strong risk factor for preterm delivery,17 and
its presence in these biologic fluids has been interpreted
as evidence of active degradation of extracellular matrix
in the membrane-decidual interface. The mechanisms in-
volved in this process are poorly understood. The pur-
pose of this study was to determine whether MMP-7 is a
physiologic constituent of amniotic fluid and if parturi-
tion, rupture of membranes, and infection may change
the amniotic fluid concentration of this enzyme.
Material and methods
Study design. A cross-sectional study was constructed
by searching our clinical database and bank of biologic
samples and included women in four main groups.
Group 1 consisted of women in the midtrimester (15-17
weeks) of pregnancy (n = 24) who underwent amnio-
centesis for genetic indications. All of these women had
normal outcomes (delivery of an appropriate-for-
gestational-age infant at term). Group 2 included pa-
tients with preterm labor and intact membranes. This
group was subdivided for study purposes into the fol-
lowing categories: (a) preterm labor who delivered at
term (n = 38), (b) preterm labor who delivered prema-
turely in the absence of microbial invasion of the amni-
otic cavity (n = 45), and (c) preterm delivery with mi-
crobial invasion of the amniotic cavity (n = 23). Preterm
labor was defined by the presence of regular uterine
contractions occurring at a minimum frequency of 2
every 10 minutes combined with documented cervical
changes in effacement and/or dilatation before 37
weeks’ gestation. Microbial invasion of the amniotic cav-
ity was defined as a positive amniotic fluid culture for
June 2000
Am J Obstet Gynecol
microorganisms. Group 3 (n = 63) consisted of women
with preterm premature rupture of membranes with (n
= 33) and without (n = 30) microbial invasion of the am-
niotic cavity. Premature rupture of the membranes was
defined as amniorrhexis before the onset of sponta-
neous labor. Membrane rupture was diagnosed with the
use of vaginal pooling, by ferning, or by a positive Ni-
trazine test. The indications for amniocentesis in
women in groups 2 and 3 were for the detection of mi-
crobial invasion of the amniotic cavity and fetal lung ma-
turity. Group 4 included women at term gestations (ie,
>37 weeks’ gestation) and was subdivided into the fol-
lowing categories: intact membranes not in labor (n =
23), intact membranes in labor (n = 22), intact mem-
branes with microbial invasion of the amniotic cavity (n
= 17), and premature rupture of membranes not in
labor with no microbial invasion of the amniotic cavity
(n = 20). Women at term not in labor underwent am-
niocentesis for assessment of fetal lung maturity before
cesarean delivery, whereas those in labor or with prema-
ture rupture of membranes underwent amniocentesis
because of labor at an uncertain gestational age or for
the diagnosis of microbial invasion of the amniotic cav-
ity. To exclude the effect of gestational age on the con-
centrations of MMP-7, the groups were matched for ges-
tational age.
Many of the samples used in this study have been used
previously in studies of amniotic fluid matrix metallopro-
teinases, cytokines, and arachidonic acid metabolites in
amniotic fluid. All women provided informed consent be-
fore collection of amniotic fluid, and the study was con-
ducted under institutional review board–approved proto-
cols.
Amniotic fluid. Amniotic fluid was collected by transab-
dominal amniocentesis in all cases. Amniotic fluid not re-
quired for clinical purposes was centrifuged for 10 min-
utes at 4°C to remove cellular and particulate matter.
Aliquots of amniotic fluid were stored at –70°C. A sample
of amniotic fluid was transported to the laboratory for
aerobic, anaerobic, and Mycoplasma species cultures ex-
cept for patients in the second trimester of pregnancy.
MMP-7 assays. Specific and sensitive enzyme-linked
immunoassays were used to determine concentrations of
MMP-7 (pro–MMP-7) in human amniotic fluid. These
assay kits (Amersham Pharmacia Biotech, Inc, Piscat-
away, NJ) are specific for quantitative determination of
human pro–MMP-7 but not the active form of MMP-7.
Moreover, this immunoassay system does not cross react
with MMP-1, MMP-2, MMP-3, MMP-9, or MT1-MMP. Be-
fore assaying the samples from this study, we validated
this assay system specifically for human amniotic fluid in
our laboratory. Validation included spike and recovery
experiments that produced parallel curves indicating
that amniotic fluid constituents did not interfere with
antigen-antibody binding in this assay system. Briefly,
Volume 182, Number 6 Maymon et al 1547
Am J Obstet Gynecol

amniotic fluid samples were incubated in duplicate wells

of the microtiter plates that had been coated with
anti–MMP-7 antibodies. During this incubation amniotic
fluid MMP-7 bound to anti–MMP-7 antibodies to form
antigen-antibody complexes. All other unbound materi-
als from the samples were removed by washing and aspi-
ration. The bound pro–MMP-7 was incubated with a per-
oxidase-labeled antigen-binding fragment antibody to
pro–MMP-7. After the incubation any excess and un-
bound material was removed by washing and aspiration.
The amount of peroxidase bound to each well was deter-
mined by the addition of ready-to-use substrate (3,3´5,5´-
tetramethylbenzidine and hydrogen peroxide in 20%
[vol/vol] dimethylformamide). The reaction was termi- A
nated by the addition of 0.1-mol/L sulfuric acid, and
the resultant color was read at 450 nm in a programma-
ble microtiter plate spectrophotometer (Ceres 900
Microplate Workstation; Bio-Tek Instruments, Inc,
Winooski, Vt). The concentrations of pro–MMP-7 in test
samples were determined by interpolation from individ-
ual standard curves (standard curve range, 0.16 to 10
ng/mL) composed of purified pro–MMP-7. The calcu-
lated interassay and intra-assay coefficients of variation
for MMP-7 immunoassay in our laboratory were 6.80%
and 6.58%, respectively. The detection limit (sensitivity)
of the assay was calculated to be 0.14 ng/mL. Amniotic
fluid samples were assayed at 1:2 dilution on the basis of
our validation data. Our spectrophotometer is capable
of measuring optical densities from 0 to 4.0. The select
B
samples that registered above the high standard were re- Fig 1. Amniotic fluid MMP-7 concentration and gestational age.
loaded and run, and the optical density fell within the A, Amniotic fluid MMP-7 concentrations were significantly
±SD curve. higher in women in second trimester than in those at term not
Statistical analysis. Nonparametric statistical tests in labor (second trimester: median, 0.54 ng/mL; range, <0.14-
(Kruskal-Wallis test, Mann-Whitney U test, Wilcoxon test 1.48 ng/mL; vs term no labor: median, 3.9 ng/mL; range, 2.1-7.8
ng/mL; P < .001, by Wilcoxon test for censored observation).
for censored observations) were used for analysis of dif- Horizontal dashed line corresponds to sensitivity of assay, 0.14
ferences between groups. The statistical packages used ng/mL. B, Amniotic fluid MMP-7 factor concentration in-
were SPSS 7.5 (SPSS Inc, Chicago, Ill) and True Epistat creased with advancing gestational age (Spearman ρ, r = –0.8;
(Epistat Services Inc, Richardson, Tex). P < .05 was con- P < .001).
sidered significant.

Results increased with advancing gestational age (Spearman ρ,

Matrilysin was detectable in 97.4% (268/275) of am-
niotic fluid samples. There were no differences in ma-
ternal age, parity, gravidity, and ethnic composition
among the study populations. The median amniotic
fluid MMP-7 concentration among patients at term not
in labor was significantly higher than the median amni-
otic fluid concentration among patients in the second
trimester of pregnancy (Fig 1, A). Similarly, a significant
correlation was found between amniotic fluid MMP-7
concentration and gestational age when the analysis in-
cluded patients with intact membranes in the second
trimester of pregnancy, those with preterm labor who
were subsequently delivered at term, and women at
term not in labor. MMP-7 amniotic fluid concentration
r = 0.8; P < .001; Fig 1, B).
Spontaneous human parturition at term was not asso-
ciated with changes in amniotic fluid concentration of
MMP-7 (Fig 2, A). In contrast, preterm parturition was as-
sociated with a significant increase in amniotic fluid con-
centration of MMP-7 (Fig 2, B). Patients with preterm
labor and intact membranes who were delivered of pre-
term neonates in the absence of microbial invasion of the
amniotic cavity) had a significantly higher median amni-
otic fluid MMP-7 than did those with preterm labor who
were delivered at term (Fig 2, B).
Spontaneous rupture of membranes in term and pre-
term gestations was not associated with significant
changes in the amniotic fluid concentration of MMP-7.
1548 Maymon et al
A
B dashed line corresponds to sensitivity of assay, 0.14 ng/mL. MIAC,
Fig 2. Amniotic fluid MMP-7 concentrations in term and pre-
term parturition. A, Term parturition was not associated with
change in amniotic fluid MMP-7 concentration. (term not in
labor: median, 3.9 ng/mL; range, 2.1-7.8 ng/mL; vs term in
labor: median, 3.3 ng/mL; range, 0.27-9.1 ng/mL; P = .9). B, Pre-
term parturition without microbial invasion of amniotic cavity
(MIAC) was associated with significant increase in amniotic fluid
MMP-7 concentration (preterm labor with preterm delivery: me-
dian, 1.7 ng/mL; range, 0.45-21.6 ng/mL; vs preterm labor with
delivery at term: median, 1.2 ng/mL; range, 0.17-42.1 ng/mL;
P < .05, by Mann-Whitney U test).
The median amniotic fluid concentration of MMP-7 was
not significantly different between patients with preterm
premature rupture of membranes and patients with pre-
term labor and intact membranes matched for gestational
age and in the absence of microbial invasion of the amni-
otic cavity (Fig 3, A). The median amniotic fluid concen-
tration of MMP-7 was lower among patients with sponta-
neous rupture of membranes at term than among patients
at term not in labor with intact membranes, although this
difference fell short of statistical significance (Fig 3, B).
Microbial invasion of the amniotic cavity was associated
with a significant increase in amniotic fluid MMP-7 con-
centration among patients with preterm labor and intact
June 2000
Am J Obstet Gynecol
A
B
Fig 3. Amniotic fluid MMP-7 concentrations among patients with
term and preterm rupture of membranes (PROM). Horizontal
Microbial invasion of amniotic cavity. A, Preterm premature rup-
ture of membranes was not associated with changes in amniotic
fluid MMP-7 concentration (preterm labor with delivery at term:
median, 1.2 ng/mL; range, 0.17-42.1 ng/mL; vs preterm prema-
ture rupture of membranes: median, 1.7 ng/mL; range, 0.29-
13.9 ng/mL; P > .05, by Mann-Whitney U test). B, Spontaneous
rupture at term was associated with significant decrease in amni-
otic fluid MMP-7 concentration compared with women at term
not in labor with intact membranes (term not in labor: median,
3.9 ng/mL; range, 2.1-7.8 ng/mL; vs term premature rupture of
membranes: median, 2.8 ng/mL; range, <0.14–13.4 ng/mL; P =
.06, by Wilcoxon test for censored observations).
membranes and among patients with preterm premature
rupture of membranes but without term labor.
Patients with preterm labor with microbial invasion of
the amniotic cavity had a significantly higher median
concentration of MMP-7 than did those with preterm
labor without infection who were delivered preterm and
those without infection who were delivered at term (Fig
4, A). Women at term with intact membranes and mi-
crobial invasion of the amniotic cavity had higher amni-
otic fluid concentrations of MMP-7 than did those at
term without microbial invasion of the amniotic cavity.
However, this difference did not reach statistical signifi-
cance (Fig 4, B). Intra-amniotic infection was associated
Volume 182, Number 6 Maymon et al 1549
Am J Obstet Gynecol

A B

C
Fig 4. Amniotic fluid MMP-7 concentration and microbial invasion of amniotic cavity (MIAC). A, Intra-amniotic infec-
tion in patients in preterm labor was associated with significant increase in amniotic fluid concentration of MMP-7. Pa-
tients with preterm labor with intra-amniotic infection had significantly higher concentration of MMP-7 than did
those with preterm labor without infection who were delivered preterm and those without infection who were deliv-
ered at term (preterm labor with microbial invasion of amniotic cavity: median, 3.2 ng/mL; range, 0.16-21.9 ng/mL;
vs preterm labor without microbial invasion of amniotic cavity and with preterm delivery: median, 1.7 ng/mL; range,
0.45-21.6 mg/mL; vs preterm labor without microbial invasion of amniotic cavity and with delivery at term: median,
1.2 ng/mL; range, 0.17-42.1 ng/mL; P < .01 for each comparison, by Kruskal-Wallis test). B, There was no significant
difference between median amniotic fluid concentrations of MMP-7 of patients at term with intact membranes and mi-
crobial invasion of amniotic cavity and patients at term without microbial invasion of amniotic cavity (microbial inva-
sion of the amniotic cavity: median, 4.1 ng/mL; range, <0.14-27.0 ng/mL; vs no microbial invasion of the amniotic cav-
ity: median, 3.3 ng/mL; range, 0.27-9.1 ng/mL; P = .5, by Wilcoxon test for censored observation). Horizontal dashed
line corresponds to sensitivity of assay, 0.14 ng/mL. C, Intra-amniotic infection in preterm premature rupture of mem-
branes (PROM) was associated with significant increase in amniotic fluid MMP-7 concentration (preterm premature
rupture of membranes without microbial invasion of amniotic cavity: median, 1.7 ng/mL; range, 0.29-13.9 ng/mL; vs
preterm premature rupture of membranes with microbial invasion of amniotic cavity: median, 3.6 ng/mL; range, 0.59-
20.3 ng/mL; P < .01, by Mann-Whitney U test).

with a significantly higher concentration of MMP-7
among women with preterm premature rupture of
membranes (Fig 4, C).
Comment
The observations reported here represent the first evi-
dence that MMP-7 is physiologically present in amniotic
fluid. MMP-7 was detected in amniotic fluid samples
from both the second and third trimesters, and its con-
centration increased with advancing gestational age. An
interesting feature of MMP-7 is that unlike other matrix
metalloproteinases, which are produced mainly by stro-
mal cells, this enzyme is produced predominantly by ep-
ithelial cells (endometrial cells, trophoblast).13, 18-20
1550 Maymon et al
Potential sources are fetal urine, amniotic cells, and resi-
dent macrophages in the amniotic cavity.21 The in-
creased concentration with gestational age suggests de-
velopmental regulation of this enzyme and its possible
participation in development, maturation, and prepara-
tion for birth.
A major finding of our study was that microbial inva-
sion of the amniotic cavity in preterm gestation was asso-
ciated with increased concentration of MMP-7 in amniotic
fluid. This was the case among both patients with intact
membranes and patients with rupture of membranes.
We propose that MMP-7 participates in extracellular ma-
trix degradation associated with intrauterine infection
by catalyzing the breakdown of fibronectin, type IV colla-
gen (which forms the basement membrane of the am-
nion), tenascin, elastin, and other extracellular matrix
components. Similar observations have been made with
gelatinase B (MMP-9)22, 23 and neutrophil collagenase
(MMP-8).24 MMP-7 is produced by macrophages21 in re-
sponse to lipopolysaccharides21 and cytokines known to
be involved in microbial invasion of the amniotic cavity25, 26
(such as interleukin 6, interleukin 1, and tumor necrosis
factor); therefore local activation of resident macro-
phages or those recruited during the course of infection
must be considered a potential source for MMP-7 in the
context of infection. This interpretation is consistent with
the observation that microbial invasion of the amniotic
cavity at term was not associated with a significant increase
in MMP-7. Indeed, concentrations of proinflammatory cy-
tokines are significantly lower among patients with micro-
bial invasion of the amniotic cavity at term than among
those with preterm labor or preterm premature rupture
of membranes.27, 28
Although spontaneous parturition at term was not as-
sociated with changes in the amniotic fluid concentration
of MMP-7, preterm labor with intact membranes leading
to preterm delivery in the absence of proven infection
was thus associated. A possible explanation for this differ-
ence is a differential role for MMP-7 in physiologic labor
(term) and pathologic labor (preterm). It is also possible
that some patients with preterm labor and negative amni-
otic fluid culture results had microbial invasion of the
amniotic cavity that escaped detection with standard mi-
crobiologic techniques.
Recently matrilysin has been implicated as a key en-
zyme in activation of α-defensin in the mouse intestine.
In knockout mice experiments in which the matrilysin
gene was deleted, null animals were more likely than the
wild species to die when fed with bacteria (Escherichia coli).
If the same relationship between MMP-7 and α-defensin
applies in humans, matrilysin may play a role in host de-
fense mechanisms in epithelial surfaces by activating an-
timicrobial peptides.29
In conclusion, MMP-7 is a physiologic constituent of
amniotic fluid, and its concentration is high in the con-
June 2000
Am J Obstet Gynecol
text of intra-amniotic infection. We propose that MMP-7
plays a role in the host defense against intra-amniotic in-
fection.
REFERENCES
1. Mecham RP, Broekelmann TJ, Fliszar CJ, Shapiro SD, Welgus
HG, Senior RM. Elastin degradation by matrix metallopro-
teinases. J Biol Chem 1997;272:18071-6.
2. Woessner JF. Matrix metalloproteinases and their inhibitors in
connective tissue remodeling. FASEB J 1991;5:2145-54.
3. Massova I, Kotra LP, Fridman R, Mobashery S. Matrix metallo-
proteinases: structures, evolution, and diversification. FASEB J
1998;12:1075-95.
4. Sellers A, Woessner JF. The extraction of a neutral metallopro-
teinase from the involuting rat uterus, and its action on cartilage
proteoglycan. Biochem J 1980;189:521-31.
5. Gaire M, Magbanua Z, McDonnell S, McNeill L, Lovett DH, Ma-
trisian LM. Structure and expression of the human gene for the
matrix metalloproteinase matrilysin. J Biol Chem 1993;269:
2032-40.
6. Woessner JF. Regulation of matrilysin in the rat uterus. Biochem
Cell Biol 1996;74:777-84.
7. Wilson CL, Matrisian LM. Matrilysin: an epithelial matrix metal-
loproteinase with potentially novel functions. Int J Biochem Cell
Biol 1996;28:123-36.
8. Woessner JF Jr. Matrilysin. Methods Enzymol 1995;248:485-95.
9. Fosang AJ, Neame PJ, Last K, Hardingham TE, Murphy G,
Hamilton JA. The interglobular domain of cartilage aggrecan is
cleaved by PUMP, gelatinases, and cathepsin B. J Biol Chem
1992;267:19470-4.
10. Patterson BC, Sang QA. Angiostatin-converting enzyme activities
of human matrilysin (MMP-7) and gelatinase B/type IV collage-
nase (MMP-9). J Biol Chem 1997;272:28823-5.
11. Marcotte PA, Kozan IM, Dorwin SA, Ryan JM. The matrix metal-
loproteinase pump-1 catalyzes formation of low molecular
weight (pro)urokinase in cultures of normal human kidney
cells. J Biol Chem 1992;267:13803-6.
12. Von Bredow DC, Nagle RB, Bowden GT, Cress AE. Cleavage of
beta 4 integrin by matrilysin. Exp Cell Res 1997;236:341-5.
13. Feng J, Woessner JF Jr, Zhu C. Matrilysin activity in the rat uterus
during the oestrous cycle and implantation. J Reprod Fertil
1998;114:347-50.
14. Wolf K, Sandner P, Kurtz A, Moll W. Messenger ribonucleic acid
levels of collagenase (MMP-13) and matrilysin (MMP-7) in vir-
gin, pregnant, and postpartum uterus and cervix of rat. En-
docrinology 1996;137:5429-34.
15. Rogers WH, Osteen KG, Matrisian LM, Navre M, Guidice LC,
Gorstein F. Expression and localization of matrilysin, a matrix
metalloproteinase, in human endometrium during the repro-
ductive cycle. Am J Obstet Gynecol 1993;168:253-60.
16. Vettraino IM, Roby J, Tolley T, Parks WC. Collagenase-I,
stromelysin-I, and matrilysin are expressed within the placenta
during multiple stages of human pregnancy. Placenta 1996;
17:557-63.
17. Goldenberg RL, Mercer BM, Iams JD, Moawad AH, Meis PJ, Das
A, et al. The Preterm Prediction Study: patterns of cervicovagi-
nal fetal fibronectin as predictors of spontaneous preterm deliv-
ery. National Institute of Child Health and Human Develop-
ment Maternal-Fetal Medicine Units Network. Am J Obstet
Gynecol 1997;177:8-12.
18. Saarialho-Kere UK, Crouch EC, Parks WC. Matrix metallopro-
teinase matrilysin is constitutively expressed in adult human ex-
ocrine epithelium. J Invest Dermatol 1995;105:190-6.
19. Busiek DF, Ross FP, McDonnell S, Murphy G, Matrisian LM, Wel-
gus HG. The matrix metalloprotease matrilysin (PUMP) is ex-
pressed in developing human mononuclear phagocytes. J Biol
Chem 1992;267:9087-92.
20. Salamonsen LA, Woolley DE. Matrix metalloproteinases and
their tissue inhibitors in endometrial remodelling and menstru-
ation. Reprod Med Rev 1996;5:185-203.
21. Busiek DF, Baragi V, Nehring LC, Parks WC, Welgus HG. Ma-
Volume 182, Number 6
Am J Obstet Gynecol
trilysin expression by human mononuclear phagocytes and its
regulation by cytokines and hormones. J Immunol 1995;154:
6484-91.
22. Athayde N, Edwin SS, Romero R, Gomez R, Maymon E, Pacora
P. A role for matrix metalloproteinase-9 in spontaneous rup-
ture of the fetal membranes. Am J Obstet Gynecol 1998;
179:1248-53.
23. Athayde N, Romero R, Gomez R, Maymon E, Pacora P, Mazor M,
et al. Matrix metalloproteinases-9 in preterm and term human
parturition. J Matern Fetal Med 1999;8:213-9.
24. Maymon E, Romero R, Pacora P, Gomez R, Athayde N, Edwin
SS, et al. Human neutrophil collagenase (matrix metallopro-
teinase) in parturition, premature rupture of the membranes,
and intrauterine infection. Am J Obstet Gynecol. In press 2000.
25. Fingleton B, McDonnell S. Cytokine regulation of matrilysin
gene expression [abstract]. Biochem Soc Trans 1997;25:155S.
26. Osteen KG, Sierra-Rivera E, Keller NR, Fox DB. Interleukin-1
alpha opposes progesterone-mediated suppression of MMP-7.
A possible role of this cytokine during human implantation. Ann
N Y Acad Sci 1997;828:137-45.
27. Romero R, Mazor M, Brandt F, Sepulveda W, Avila C, Cotton DB,
Dinarello CA. Interleukin-1 alpha and interleukin-1 beta in pre-
term and term human parturition. Am J Reprod Immunol
1992;27:117-23.
28. Romero R, Mazor M, Sepulveda W, Avila C, Copeland D,
Williams J. Tumor necrosis factor in preterm and term labor. Am
J Obstet Gynecol 1992;166:1576-87.
29. Wilson CL, Ouellette AJ, Satchell DP, Ayabe T, López-Boado YS,
Stratman JL, et al. Regulation of intestinal alpha-defensin activa-
tion by the metalloproteinase matrilysin in innate host defense.
Science 1999;286:113-7.
Discussion
DR JAMES A. MCGREGOR, Denver, Colorado. Maymon
and colleagues from the National Institute of Child
Health and Human Development’s Perinatology Re-
search Branch have once again extended our knowledge
of the pathobiologic characteristics of preterm labor and
rupture of the membranes, as well as physiologic mecha-
nisms of human parturition. This carefully performed
cross-sectional study and chemical analysis of the Perina-
tology Research Branch’s uniquely large and well-charac-
terized collection of biologic samples shows that ma-
trilysin (pro–MMP-7) is both a constitutive and an
inducible product of fetal-trophoblastic or maternal tis-
sues that plays a role in normal gestational growth and re-
modeling of fetal membranes and in pathologic
processes involved in infection- or inflammation-medi-
ated preterm parturition. The authors were careful to
note that preterm labor without culturable microbes in
amniotic fluid was also associated with an increased am-
niotic fluid MMP-7 concentration. Of course, fetal or tro-
phoblastic inflammation may be present without culture-
proven amniotic fluid invasion (that is, microbes present
below culture-threshold levels or containment of mi-
crobes within amnion-chorion, placenta, the fetus, or ma-
ternal decidua).
This investigation and others, including those by For-
tunato et al,1 Draper et al,2 Dudley,3 and Majzoub et al,4
illustrate how host molecular mechanisms involved with
growth and development, inflammation, and even stress
responsiveness are now understood to be fully integrated
and reinforcing in producing labor. It has already been
speculated that we are approaching a standard model or
a central theory of parturition.4 The clinical problems of
Maymon et al 1551
prematurity can now be seen as a special case that vari-
ously features infection or inflammation, bleeding, hy-
poxemia, overdistention, and endocrine or paracrine
and apoptotic mechanisms. Each of these processes is
controlled by genomic mechanisms (complementary de-
oxyribonucleic acid, messenger ribonucleic acid), which
in turn control cellular, tissue, organ, and multiorgan
processes in the fetus and the mother.
The examples of the authors of this biologic inte-
grated responsiveness include observations that MMP-7
(1) cleaves fibronectin and other structural proteins,
(2) activates cytokines, including tumor necrosis factor
α, and (3) induces defensins (innate host defense an-
tibiotic-like peptides). New observations suggest that the
father plays a more important role than is commonly
credited. Paternal genes are differentially expressed in
the placenta, which is called placental imprinting. Of
course microbial enzymes can destroy gestational tissues
directly.
In nonpregnant life and in nonterm pregnancy, dy-
namic and potentially destructive proteases such as MMP-
7 are closely controlled by opposing homeostatic mecha-
nisms, such as production of anti–MMP-7 or tissue
inhibitors of metalloproteinases. During late pregnancy
these processes appear teleologically “designed” to cause
parturition and bring about birth (a change of state
called allostasis). The fact that potentially preventable dis-
ease processes, such as intrauterine infection, activate cell
mechanisms involved with growth and development, con-
nective tissue remodeling, innate host defenses, cortico-
tropin-releasing hormone, and placental clock mecha-
nisms, as well as with inflammation, reinforces the notion
that steps taken to prevent prematurity must be taken
during early pregnancy—or even before conception—
and that these steps must be comprehensive and clini-
cally integrated. I am confident that the ongoing contri-
butions of Maymon and colleagues will continue to
enhance our ability to provide science-based obstetric
care.
REFERENCES
1. Fortunato SJ, Menon R, Lombardi SJ. Interleukin-10 and trans-
forming growth factor-beta inhibit amniochorion tumor necro-
sis factor-alpha production by contrasting mechanisms of action:
therapeutic implications in prematurity. Am J Obstet Gynecol
1997;177:803-9.
2. Draper D, Hall J, Jones W, Beutz M, Heine RP, Porecco R, et al.
Elevated protease activities in human amnion and chorion cor-
relate with preterm premature rupture of membrane. Am J Ob-
stet Gynecol 1995;1173:1506-12.
3. Dudley DJ. Immunoendocrinology of preterm labor: the link be-
tween corticotropin-releasing hormone and inflammation. Am J
Obstet Gynecol 1999;180(1 Pt 3):S251-6.
4. Majzoub JA, McGregor JA, Lockwood CJ, Smith R, Taggart MS,
Schulkin J. A central theory of preterm and term labor: putative
role for corticotropin-releasing hormone. Am J Obstet Gynecol
1999;180:(1 Pt 3):S232-41.
DR STEPHEN FORTUNATO, Nashville, Tennessee. I think
that it is important, as we look at this area, to realize that
the inhibition and regulation of the family of matrix me-
talloproteinases is an incredibly complex area. Not only
are the matrix metalloproteinases controlled at the gene
1552 Maymon et al
level, they also have activators and inhibitors and are
translationally controlled. When we look at which matrix
metalloproteinases are truly active in a particular area, we
really cannot look for a single matrix metalloproteinase
that has a certain function, for instance, in infection or
premature rupture of membranes, because the matrix
metalloproteinases probably all act together in a cohesive
symphony.
I think that matrilysin is important. As the authors
noted, it lacks the hemopexin domain, so it does not bind
as well to the large molecules and is not as active in de-
grading some of these molecules. However, it is most
likely very active in activating some of the other matrix
metalloproteinases that play a role in this process and,
again as the authors said, in aiding the body’s defense
against infections.
DR PAT COLLINS, Maywood, Illinois. I have a question
about the antibody and its identification. Dr Maymon,
you were saying that the test measured pro–MMP-7 and
that there were no differences between patients with rup-
tured membranes and labor. Do you think that might be
because you are measuring the pro-form rather than the
active enzyme form?
Next, even though there were differences that were
statistically significant, the concentration differences
were actually fairly small in terms of nanograms per milli-
liter. Would it not be better to be able to measure an en-
zyme activity in units per milligram or a similar unit? That
is really the information that you want.
DR MAYMON (Closing). First, I thank Dr McGregor,
who is a pioneer in the field of degrading enzymes. Ten
years ago he suggested that human neutrophils and
their constituent enzymes may act in concert with bacte-
ria and their protease(s) in weakening amniochorion1
and may possibly predispose to premature rupture of
membranes. Subsequently, in an American Journal of Ob-
stetrics and Gynecology publication, he and his colleagues2
reported elevated protease activity in human amnion
and chorion in patients with preterm premature rup-
ture of membranes.
I will start with the question from Dr Collins regarding
the activity of the enzyme. The assay that was used in this
study is an enzyme-linked immunosorbent assay that mea-
sures only the proform, not the active forms, of MMP-7.
As such we were not able to draw conclusions regarding
the activity of the enzyme. An important question is
whether changes in immunoreactive MMP-7 were re-
flected by changes in enzymatic activity. Since the submis-
sion of our article we have continued our studies and
have conducted zymography on samples of amniotic fluid
from different groups of patients included in this study.
For those who are not familiar with zymography, this is
a simple assay to detect enzymatic activity. For example,
the substrate we used in zymography is casein, a protein
known to be degraded by matrilysin. Amniotic fluid sam-
ples from patients with preterm labor with negative amni-
otic fluid culture results who delivered at term, those with
negative culture results who delivered preterm, and those
with positive amniotic fluid culture results were run.
June 2000
Am J Obstet Gynecol
It is clear that active forms of the enzyme are de-
tectable in patients with intra-amniotic infection. These
data thus suggest that patients with increased concentra-
tions of immunoreactive enzyme may have increased ac-
tivity of the enzyme in vivo. A much larger study is re-
quired to estimate enzymatic activity in amniotic fluid.
However, the limitations of zymography as a quantitative
technique are well known.
The cellular sources of matrilysin in amniotic fluid
were not a subject of our study, but this area is currently
being investigated by our group. The identification of
these sources is extremely important, and I will share with
you some data that address the question. We have used
two approaches–first, to examine whether fetal mem-
branes have immunoreactive MMP-7 and, second, to ex-
plore fetal sources for this enzyme.
We have performed immunohistochemical studies to
determine expression of matrilysin in fetal membranes.
We used frozen tissues stained with a monoclonal anti-
body against MMP-7.
Amniotic cells in fetal membranes from a patient with-
out intra-amniotic infection expressed MMP-7. The con-
nective tissue underneath the amnion showed no expres-
sion, and most chorionic cells show no staining for
MMP-7. In contrast, in a patient with documented intra-
amniotic infection, there was an intense staining of cells
in the chorion for MMP-7.
Therefore these studies suggest that amniotic cells
express MMP-7 and that this may be a source of amni-
otic fluid MMP-7. However, our observations of in-
creased expression of MMP-7 in the membranes and
specifically in the chorion in cases of infection indicate
that trophoblast cells are a potential cell of origin. This
observation has relevance, because our group previously
demonstrated that chorionic cells can express α-de-
fensins, which are antimicrobial peptides. In addition,
in 1993 Rodgers et al3 showed with in situ hybridization
and immunohistochemical methods that MMP-7 is pre-
sent in human endometrium. To explore whether the
human fetus could be a source of amniotic fluid
matrilysin, we assayed matrilysin in a few samples of first-
voided neonatal urine. In approximately 50% of sam-
ples immunoreactive pro–MMP-7 was detected. There-
fore the human fetus is a potential source of amniotic
fluid MMP-7.
Regarding the question of whether serial samples were
taken from the same patient, the answer is no.
There were no differences in demographic data
among the different groups. The patients in the different
groups were matched for gestational age. Genital My-
coplasma species (specifically Ureaplasma urealyticum) was
the single most common microorganism isolated from
the fluid. We have examined our data and are not able to
discern a relationship between amniotic fluid concentra-
tions of MMP-7 and the specific microorganisms. Deter-
mining whether some microorganisms may be more po-
tent than others in stimulating the production or release
of MMP-7 would require in vitro studies, which have not
been performed to date.
Volume 182, Number 6
Am J Obstet Gynecol
I am reluctant to make any comments about the clini-
cal implication of clinical work, because this is the first re-
port of increased amniotic fluid MMP-7 concentration
among patients with intra-amniotic infection. Matrilysin
was implicated as a key enzyme in the activation of α-de-
fensin in the mouse intestine in an article recently pub-
lished in Science.4 In knockout experiments in which the
matrilysin gene was deleted, null animals (without the
MMP7 gene) were more likely to die when fed with bac-
teria, specifically E coli, than were the wild-type animals. If
the same relationship between MMP-7 and α-defensins
applies to human beings, matrilysin may play a role in
host defense mechanisms in epithelial surfaces by activat-
ing antimicrobial peptides.
Maymon et al 1553
REFERENCES
1. McGregor JA, Schoonmaker JN, Lunt BD, Lawellin DW. Antibi-
otic inhibition of bacterially induced fetal membrane weaken-
ing. Obstet Gynecol 1990;76:124-8.
2. Draper D, McGregor J, Hall J, Jones W, Beutz M, Heine RP, et al.
Elevated protease activities in human amnion and chorion cor-
relate with preterm premature rupture of membranes. Am J Ob-
stet Gynecol 1995;173:1506-12.
3. Rogers WH, Osteen KG, Matrisian LM, Navre M, Guidice LC,
Gorstein F. Expression and localization of matrilysin, a matrix
metalloproteinase, in human endometrium during the repro-
ductive cycle. Am J Obstet Gynecol 1993;168:253-60.
4. Wilson CL, Ouellette AJ, Satchell DP, Ayabe T, López-Boado YS,
Stratman JL, et al. Regulation of intestinal α-defensin activation
by the metalloproteinase matrilysin in innate host defense. Sci-
ence 1999;286:113-7.