Reduction-oxidation (redox) state regulation of matrix

metalloproteinase activity in human fetal membranes

Irina A. Buhimschi, MD, Wayne B. Kramer, MD, Catalin S. Buhimschi, MD, Loren P. Thompson,
PhD, and Carl P. Weiner, MD
Baltimore, Maryland

OBJECTIVE: The mechanisms underlying membrane rupture at term and preterm are obscure.
Collagenolytic activity of matrix metalloproteinases in amniochorionic membranes increases during sponta-
neous term and preterm labor associated with intra-amniotic infection. We sought to test the hypothesis that
reduction-oxidation homeostasis, which is altered in inflammatory states, directly regulates amniochorionic
matrix metalloproteinases.
STUDY DESIGN: Membranes were collected from 7 patients undergoing elective cesarean delivery at term,
rinsed thoroughly, and immediately incubated in phosphate-buffered sodium chloride solution at 37C for 24
hours. Matrix metalloproteinase activity in the culture medium was assayed by substrate-gel electrophoresis
and normalized against the dry weight of the tissue incubated. Superoxide anions were generated in the
presence of membranes by a xanthine (2 mmol/L) and xanthine oxidase (20 mU/mL) mixture and monitored
by reduction of ferricytochrome c to ferrocytochrome c. Incubations were performed in the presence of
xanthine alone, a xanthinexanthine oxidase mixture, superoxide dismutase (500 U/mL), a xanthinexan-
thine oxidasesuperoxide dismutase mixture, nitro-L-arginine (a nitric oxide synthase inhibitor, 1 mmol/L),
xanthinexanthine oxidasenitro-L-arginine, S-nitroso-N-acetylpenicillamine (a nitric oxide donor, 10 mmol/L),
xanthinexanthine oxidaseS-nitroso-N-acetylpenicillamine, N-acetylcysteine (a thiol-containing antioxidant,
0.1, 1, or 10 mmol/L), lipopolysaccharide (100 ng/mL), or lipopolysaccharideN-acetylcysteine. Intracellular
generation of superoxide anions was monitored by the reduction of nitroblue tetrazolium to formazan.
RESULTS: Basal matrix metalloproteinase 9 and matrix metalloproteinase 2 levels were detected in all sam-
ples. Superoxide anions significantly increased matrix metalloproteinase 9 activity but did not increase matrix
metalloproteinase 2 activity, which effect was reversed by the addition of superoxide dismutase.
N-acetylcysteine reduced basal activity of both matrix metalloproteinase 9 and matrix metalloproteinase 2
to 20%. Importantly, N-acetylcysteine completely inhibited intracellular formazan formation in cultured mem-
branes both in the absence and in the presence of lipopolysaccharide. Neither nitric oxide synthase inhibition
nor the nitric oxide donor S-nitroso-N-acetylpenicillamine had any effect on fetal membrane matrix metallo-
proteinase activity.
CONCLUSION: Matrix metalloproteinase activity in human fetal membranes is reduction-oxidation (redox)
regulated. Matrix metalloproteinase 9 activity in human fetal membranes is directly increased by superoxide
anion, a byproduct of macrophages and neutrophils. Neither nitric oxide donors nor nitric oxide synthase in-
hibitors significantly affect matrix metalloproteinase activity in human fetal membranes. The glutathione pre-
cursor N-acetylcysteine dramatically inhibits amniochorionic matrix metalloproteinase activity in addition to
inhibiting intrinsic superoxide generation within the tissue. Thus thiol-reducing agents, such as N-acetylcys-
teine, may be beneficial in preventing preterm premature rupture of the membranes. (Am J Obstet Gynecol
2000;182:458-64.)

Key words: N-acetylcysteine, superoxide, amnion, chorion, nitric oxide

Preterm birth is the major cause of perinatal morbidity
and mortality in the world. Prematurity is responsible for
75% of infant deaths and 50% of the long-term neuro-
From the Division of Perinatal Research, Department of Obstetrics,
Gynecology, and Reproductive Sciences, University of Maryland School
of Medicine.
Received for publication May 20, 1999; revised September 1, 1999;
accepted October 5, 1999.
Reprint requests: Irina Buhimschi, MD, Department of Obstetrics,
Gynecology, and Reproductive Sciences, University of Maryland School
of Medicine, F.C. Bressler Research Laboratories 11-011, 655 W
Baltimore St, Baltimore, MD 21201-1559.
Copyright 2000 by Mosby, Inc.
0002-9378/2000 $12.00 + 0 6/1/103514
logic handicaps, including cerebral palsy, blindness, deaf-
ness, and developmental delays.1 The survival rate of
neonates is improved by 2% per day from the 23rd to the
26th week of pregnancy (ie, from 16% at 23 weeks to
57% at 26 weeks), reaching 80% survival at 28 weeks and
>90% by 30 weeks of gestation.2 Therefore any treatment
that prevents premature birth will profoundly reduce
neonatal mortality and morbidity rates. Although the
cause of most preterm births remains unknown, current
evidence suggests a multifactorial origin of preterm
labor, with an inflammatory syndrome playing a large
role.3 Microbial invasion of the amniotic cavity occurs in
10% of the patients with preterm labor and intact mem-

458
Volume 182, Number 2
Am J Obstet Gynecol
branes and in 38% of the patients with preterm prema-
ture rupture of membranes.3 Techniques such as poly-
merase chain reaction, which are more sensitive than cul-
ture, detect bacteria in 60% of the pregnancies
complicated with preterm labor.4 Infection-related
preterm labor most likely involves the release of inflam-
matory cytokines by host defense mechanisms in re-
sponse to bacterial products (ie, lipopolysaccharide). It is
believed that the proinflammatory cytokines (eg, inter-
leukins: IL-1, IL-6, and IL-8 and tumor necrosis factor a
[TNF-a]) stimulate the production of uterotonins, such
as prostaglandins, leukotrienes, and oxytocin, in the de-
cidua and fetal membranes, eventually leading to the
onset of labor.3 The key event of any inflammatory
process is the recruitment of inflammatory cells, such as
neutrophils and macrophages, which in turn trigger the
release of cytokines and free radicals (nitric oxide and su-
peroxide radical) and activate metalloproteinases with
subsequent degradation of the extracellular matrix, lead-
ing, at least in the case of infection-related preterm labor,
to effacement and dilation of the uterine cervix or to the
rupture of fetal membranes.3
Matrix metalloproteinases are a family of endopepti-
dases that collectively cleave most, if not all, the con-
stituents of the extracellular matrix. Major members of
this family are interstitial collagenase (MMP-1), 72-kd type
IV collagenase (MMP-2, 72-kd gelatinase), stromelysin
(MMP-3), and 92-kd type IV collagenase (MMP-9, 92-kd
gelatinase). The enzymes are secreted into the intercellu-
lar compartment as proMMPs and require an activating
agent either to cleave a propeptide sequence and/or to
perturb their conformation. Autocatalytic processes, with
further propeptide sequence cleavage, result in the fully
active enzyme.5
Matrix metalloproteinases are regulated by binding to
high-affinity tissue inhibitors of metalloproteinase. Several
matrix metalloproteinases (MMP-1, MMP-2, MMP-3, and
MMP-9), as well as tissue inhibitors of MMP-1 and MMP-2,
are expressed in fetal membranes.6, 7 Increased MMP-9 ac-
tivity in amniotic fluid,8 human fetal membranes,6 and
human plasma9 is reported during spontaneous labor at
term. However, recent data suggest that both preterm
labor and preterm premature rupture of the membranes
are associated with further elevated activity of MMP-9 in
amniotic fluid8 and amniochorionic membranes.6
Matrix metalloproteinase (particularly MMP-9) activa-
tion is a general feature in several inflammatory
processes with high cytokine output, such as periodontal
disease, rheumatoid arthritis, and asthmatic airway in-
flammation.5 Furthermore, data from several in vitro ex-
periments in culture conditions have demonstrated a
causative relationship among multiple cytokines (IL-1,
IL-6, IL-8, TNF-a), lipopolysaccharide (LPS), and MMP-9
expression or activity.6, 10, 11 In addition, modulation of
the reduction-oxidation state of the environment has
Buhimschi et al 459
been shown to alter matrix metalloproteinase activity di-
rectly, as well as the magnitude of the response induced
by cytokines.12, 13 Specifically, an increase in pro-oxidants
or a decrease in antioxidants (ie, altering the reduction-
oxidation balance) increases matrix metalloproteinase
activity.
During normal respiration, a small but significant
amount (1%-5%) of the total oxygen consumed generates
partially reduced intermediates (reactive oxygen species),
such as the superoxide radical anion, hydrogen peroxide,
and the hydroxyl radical. These partially reduced inter-
mediates are very reactive and are the cause of oxygen
toxicity and mutagenicity. Most living organisms have
evolved well-integrated antioxidant defense mechanisms
(scavengers) that include superoxide dismutases, catalase,
glutathione peroxidases, reduced glutathione, b-caro-
tene, and vitamins C (ascorbic acid) and E.
Although a link between matrix metalloproteinase ac-
tivity and reactive oxygen species in amniotic membranes
has not been made to date, there is evidence supporting
this hypothesis in other experimental models.
Incubation of cultured human vascular smooth muscle
cells with a superoxide-generating mixture increases the
amount of active matrix metalloproteinases.12 Further-
more, antioxidants, such as N-acetyl cysteine (a mem-
brane-permeable glutathione precursor), inhibit the
chondrolytic activity in cultured cartilage of fibronectin
fragments14 known to act through catabolic cytokine ac-
tion mediated by the interleukins IL-1 and IL-6 and by
TNF-a.15
We hypothesize that reduction-oxidation homeostasis,
known to be altered in inflammatory states, directly regu-
lates amniochorionic matrix metalloproteinase activity.
Therefore the aim of this study was to investigate whether
local changes in levels of reactive oxygen species (ie, in-
creased superoxide radical anion or increased glu-
tathione) correlate with matrix metalloproteinase activity
in human fetal membranes.
Material and methods
Tissues. Full-thickness amniochorionic membranes
were collected from 7 patients without complications of
pregnancy undergoing elective cesarean delivery at term.
The fragment collected was selected from a region at least
10 cm away from the placenta. Immediately after collec-
tion, membranes were placed in minimum essential
medium (Gibco, Grand Island, NY) on ice. Tissue samples
were then cut under sterile conditions with a scalpel blade
into smaller pieces of similar size. Over the next 30 min-
utes, the pieces of membranes were rinsed thoroughly
with several changes of minimum essential medium and
then sterile phosphate-buffered sodium chloride solution
(pH 7.8). Preweighed amounts of tissue (approximately
500 mg) consisting of several pieces selected at random
were placed in a carbon dioxide incubator with a humidi-
460 Buhimschi et al
fied chamber at 37C for 24 hours. Incubations were car-
ried out in a total volume of 2 mL of buffer. Both wet
weight and dry weight of tissue pieces were determined
after incubation and after drying overnight at 60C, re-
spectively. After incubation, the culture-conditioned
medium was collected, briefly spun at 1000g to remove
cellular debris, and loaded directly onto gelatin gels. The
concentration of protein in the aliquots of medium was
measured with a BCA (Pierce, Rockford, Ill) kit.
Drugs and incubations. Incubations were performed in
the presence of xanthine (2 mmol/L), xanthine oxidase
(20 mU/mL), the combination of xanthine plus xan-
thine oxidase, superoxide dismutase (500 U/mL), xan-
thinexanthine oxidasesuperoxide dismutase, nitro-L-
arginine (a nonspecific nitric oxide synthase inhibitor, 1
mmol/L), xanthinexanthine oxidasenitro-L-arginine,
S-nitroso-N-acetylpenicillamine (a nitric oxide donor, 10
mmol/L), xanthinexanthine oxidaseS-nitroso-N-acetyl-
penicillamine, N-acetylcysteine (0.1, 1, and 10 mmol/L),
and lipopolysaccharide (100 ng/mL). All chemicals were
purchased from Sigma (St Louis, Mo) unless otherwise
specified.
Extracellular superoxide anion generation. Superoxide
anion was generated enzymatically from the reaction be-
tween 2-mmol/L xanthine and 20 mU/mL xanthine oxi-
dase, as described by McCord and Fridovich.16 Although
these amounts are larger than those routinely used for in
vitro generation of superoxide anion in cell-free experi-
ments, we have based our decision on the presence of
both xanthine oxidase and superoxide dismutase in fetal
membranes17 and on our assumption that the expression
of these enzymes will change the efficiency of the xan-
thine plus xanthine oxidase reaction. As a result, we
monitored the efficiency of the xanthine plus xanthine
oxidase reaction in our experimental setting by use of a
cytochrome c assay both in tissue-free conditions and in
the presence of the tissue.18 The reduction of ferri
cytochrome c to ferrocytochrome c is a process highly
selective for extracellular superoxide anion and associ-
ated with a change in absorbance at 550 nm within 15
minutes. As proof for the specificity of the reaction, su-
peroxide dismutase (500 U/mL) completely inhibited
the change in absorbance at 550 nm.
Intracellular superoxide anion generation. Pieces of
fetal membranes were immersed in Krebs bicarbonate
buffer (sodium chloride, 118 mmol/L; potassium chlo-
ride, 4.7 mmol/L; magnesium sulfate, 1.18 mmol/L;
monobasic potassium phosphate, 1.18 mmol/L; D-glu-
cose, 11.1 mmol/L; ethylenediaminetetraacetic acid,
0.016 mmol/L; calcium chloride, 2.2 mmol/L; sodium bi-
carbonate, 15.8 mmol/L [pH 7.35-7.4]) containing 1
mg/mL nitroblue tetrazolium for 180 minutes in the car-
bon dioxide incubator with humidified chamber at 37C.
The soluble yellow form of nitroblue tetrazolium is re-
duced by the superoxide radical anion generated within
February 2000
Am J Obstet Gynecol
the tissue to form an insoluble precipitate of blue-purple
formazan. After incubation, the fetal membranes were
washed in sodium chloride solution, photographed,
placed in 4% formaldehyde, and kept overnight in a re-
frigerator. Specimens were then embedded in OCT
Tissue Tek (Fisher Scientific, Pittsburgh, Pa) embedding
medium and frozen in liquid nitrogen. The frozen speci-
mens were cut into 20-m m sections with a cryostat,
mounted onto electrostatically treated glass slides
(Fisherbrand Superfrost Plus; Fisher Scientific), and
counterstained with neutral red stain. They were immedi-
ately investigated by light microscopy with a Nikon Eclipse
E1000M microscope connected to an MTI DC330 3CCD
Color Camera (Dage MTI Inc, Michigan City, Mich).
Sodium dodecyl sulfatesubstrate gel electrophoresis (zy-
mography). For detection of matrix metalloproteinases,
equal volumes (15 m L) of conditioned medium (contain-
ing the activated matrix metalloproteinases) were mixed
in a 1:3 ratio with substrate gel buffer [10% sodium dode-
cyl sulfate, 4% sucrose, 0.25-mol/L tris(hydroxy-
methyl)aminomethane (Tris) hydrochloride (pH 6.8),
and 0.1% bromophenol blue] and loaded directly onto
gels without boiling (as in reference 12 with modifica-
tions). Briefly, type I gelatin was added to a standard
Laemmli 10% acrylamide polymerization mixture at a
final concentration of 1 mg/mL. After electrophoresis, the
proteins in the gel were renatured by exchanging sodium
dodecyl sulfate with Triton X-100, and the gels were incu-
bated overnight at 37C in 50-mmol/L Tris hydrochloride
(pH 7.4) containing 10-mmol/L calcium chloride and
0.05% Brij 35. At the end of the incubation, the gels were
stained with 0.5% Coomassie blue R-250 (Bio-Rad,
Hercules, Calif) in 40% methanol10% acetic acid and
then destained in 40% methanol10% acetic acid for 1
hour. Proteins having gelatinolytic activity appeared as dis-
crete translucent areas of lytic activity on an otherwise blue
gel. When the gels are incubated in parallel with 0.01-
mol/L ethylenediaminetetraacetic acid, disappearance of
lytic bands confirms the metal dependence of matrix met-
alloproteinase activity. Migration of proteins was com-
pared with that of prestained broad molecular weight
markers (Bio-Rad). Furthermore, zymography standards
of 10 ng/lane (recombinant MMP-2/MMP-9; Calbiochem,
La Jolla, Calif) were run along with the samples.
To quantify the gelatinolytic activity of matrix metallo-
proteinases, the wet gel was scanned on a GS-670 Image
Densitometer (Bio-Rad) and analyzed by Multianalyst
v1.1 (Bio-Rad) volume quantification system software.
After image inversion and background subtraction, the
amount of lytic activity, which now appears as a black
band on a white background, is estimated as optical den-
sity over the area of the band. These values were normal-
ized against the dry weight of the tissue incubated and
further normalized against the value from the lane where
the tissue was incubated in phosphate-buffered sodium
Volume 182, Number 2 Buhimschi et al 461
Am J Obstet Gynecol

A B
Fig 1. MMP-2 (A) and MMP-9 (B) activity in fetal membranes incubated in presence of phosphate-buffered sodium
chloride solution only (CRL), xanthine (X, 2 mmol/L), superoxide dismutase (SOD, 500 U/mL), xanthinexanthine
oxidase (X+XO), or xanthinexanthine oxidasesuperoxide dismutase (X+XO+SOD). Matrix metalloproteinase levels
were quantified and normalized as described in the Methods section. Data are presented as mean + SEM from 7 dif-
ferent patients and analyzed by one-way analysis of variance followed by post hoc comparisons with multiple Tukey
tests. Means with at least one common letter designation are not different, at P > .05.

chloride solution alone on the same gel. To evaluate the
reliability of the image-analysis system, differing amounts
of zymography standard (5, 10, and 20 ng/lane) were
loaded onto one gel, and MMP-2 (72 kd) and MMP-9 (92
kd) activities were quantified separately. The correlation
coefficient between matrix metalloproteinase concentra-
tion as quantified by substrate gel analysis and the
amount of protein loaded approached 1. To evaluate the
reliability of the normalization, differing amounts of am-
niochorionic tissue (100, 250, 500, and 750 mg) were in-
cubated, and MMP-2 and MMP-9 activities were plotted
against the dry weight of the tissue. The correlation coef-
ficient was 0.98.
Statistical analysis. The results are presented as mean +
1 SEM of the percentage activity from the lane incubated
in phosphate-buffered sodium chloride solution (con-
trol) from the same patient and run in the same substrate
gel electrophoresis. Multiple comparisons between
groups were performed by one-way analysis of variance
followed by multiple post-hoc Tukey tests. A P value < .05
was considered as the limit for statistical significance.
Results
Matrix metalloproteinase activity in human term am-
niochorionic membranes in the absence and presence of
reduction-oxidation balance modulators. MMP-9 and
MMP-2 corresponding to the 92- and 72-kd lytic bands,
respectively, were detected in all tissues, at least after cul-
ture conditions (Fig 1). In some of the patients an addi-
tional band corresponding to proMMP-9 was observed.
There was a large variability in the intensity of the lytic
bands among the 7 patients, justifying the need for the
normalization described in the Methods section. The su-
peroxide anion generated by the xanthinexanthine oxi-
dase reaction induced a nonsignificant increase in the
total amount of protein released in the conditioned
medium, also justifying the need for a normalization
against the dry weight of the tissue rather than against
the protein concentration in the medium. The efficiency
of the xanthinexanthine oxidase reaction was con-
firmed for each experiment by means of the cytochrome
C reduction assay.
As illustrated in Fig 1, incubation with xanthinexan-
thine oxidase significantly increased MMP-9 levels. This in-
crease was reversed by simultaneous addition of superox-
ide dismutase. A small increase in MMP-9 was observed in
some of the patients with xanthine alone, probably reflect-
ing endogenous xanthine oxidase activity. Incubation with
10-mmol/L N-acetylcysteine dramatically reduced both
MMP-2 and MMP-9 activities to 20% of the levels from the
control tissue (Fig 2). N-Acetylcysteine also prevented the
increase in MMP-9 induced by xanthinexanthine oxidase
(Fig 2, B). Neither nitric oxide synthase inhibition nor the
nitric oxide donor S-nitroso-N-acetyl penicillamine had
any significant effect on fetal membrane matrix metallo-
proteinase activity (data not shown). A high variability in
the response after nitro-L-arginine was noted.
Fig 3 illustrates the gelatinolytic activity in fetal mem-
branes in the presence or absence of lipopolysaccharide.
Zymographically, we did not observe a significant in-
crease in matrix metalloproteinase activity after
lipopolysaccharide alone. However, N-acetylcysteine (10
mmol/L) dramatically inhibited MMP-2 and MMP-9
both in the presence and in the absence of lipopolysac-
charide. In some of the experiments with low general ma-
trix metalloproteinase concentration, 10-mmol/L N-
acetylcysteine completely inhibited any gelatinolytic
activity in the amniochorion.
Intracellular generation of superoxide anion in fetal
membranes. After exposure to nitroblue tetrazolium over
the 3-hour incubation period, there was spontaneous for-
mation of formazan detectable macroscopically. Exposure
462 Buhimschi et al February 2000
Am J Obstet Gynecol

A B
Fig 2. MMP-2 (A) and MMP-9 (B) activity in fetal membranes incubated in presence of phosphate-buffered sodium
chloride solution only (CRL), N-acetylcysteine (NAC10, 10 mmol/L), xanthinexanthine oxidase (X+XO), or xan-
thinexanthine oxidaseN-acetylcysteine (X+XO+NAC10). Matrix metalloproteinase levels were quantified and nor-
malized as described in the Methods section. Data are presented as mean + SEM from 7 different patients and ana-
lyzed by one-way analysis of variance followed by post hoc comparisons with multiple Tukey tests. Means with at least
one common letter designation are not different, at P > .05.

A B
Fig 3. MMP-2 (A) and MMP-9 (B) activity in fetal membranes incubated in presence of phosphate-buffered sodium
chloride solution only (CRL), N-acetylcysteine (NAC10, 10 mmol/L), lipopolysaccharide (LPS, 100 nmol/L), or
lipopolysaccharideN-acetylcysteine (LPS+NAC10). Matrix metalloproteinase levels were quantified and normalized as
described in the Methods section. Data are presented as mean + SEM from 5 different patients and analyzed by one-
way analysis of variance followed by post hoc comparisons with multiple Tukey tests. Means with different letter designa-
tions are different, at P < .05.

to the xanthinexanthine oxidase reaction produced
both intracellular (intense purple discoloration of the tis-
sue) and extracellular (discoloration of the incubation
medium) formazan formation. The addition of superox-
ide dismutase completely inhibited the extracellular and
only partially inhibited the intracellular formazan forma-
tion. It is interesting that lipopolysaccharide administered
within either 3 or 24 hours did not increase superoxide
anion production over the levels seen in tissues incubated
with Krebs buffer alone. N-acetylcysteine (10 mmol/L)
completely abolished formazan formation in the incu-
bated tissue. On examination by light microscopy, the de-
posits of formazan induced by xanthinexanthine oxidase
were localized mainly within the chorion laeve, decidua,
and amniotic epithelium (Fig 4, B). Scattered deposits
were also seen in the connective tissue layer between the
amnion and chorion. In control tissues discrete deposits
were localized in the chorion and the decidua, as well as
in large areas of the amniotic epithelium (Fig 4, A). In tis-
sues exposed to N-acetylcysteine (Fig 4, C), there were no
formazan deposits, indicating a complete inhibition of in-
tracellular superoxide anion formation.
Comment
This study demonstrates (1) that MMP-9 levels in
human fetal membranes are directly increased by super-
oxide anion, (2) that the glutathione precursor N-acetyl-
cysteine dramatically inhibits amniochorionic matrix me-
talloproteinase activity in addition to the inhibition of
intrinsic superoxide generation within the tissue, and (3)
that neither nitric oxide nor nitric oxide synthase in-
hibitors significantly affect matrix metalloproteinase activ-
ity in human fetal membranes. Together, these findings
suggest that the overall reduction-oxidation status of the
Volume 182, Number 2
Am J Obstet Gynecol
Fig 4. Microscopic aspect of fetal membranes incubated for 3 hours in Krebsnitroblue tetrazolium buffer (A) and in
presence of xanthinexanthine oxidase (B) or 10-mmol/L N-acetylcysteine (C). (Original magnification 150. Inset,
Original magnification 1500.)
environment may be an important modulator of matrix
metalloproteinase levels. An increase in the oxidative
state could enhance extracellular matrix degradation.
Conversely, thiol-reducing agents, such as N-acetylcys-
teine, could act as inhibitors of matrix metalloproteinase
activation within fetal membranes and prevent premature
rupture of membranes caused by inflammation.
Changes in the oxidative environment have previously
been shown to regulate matrix metalloproteinase activity
in several cell types. Human heart fibroblasts are reduc-
tion-oxidationsensitive and under oxidative conditions
are activated to concurrently express metalloproteinases
and tissue inhibitors of metalloproteinase.19 Further-
more, thiol but not nonthiol (reduced glutathione and
N-acetylcysteine) reducing agents inhibit matrix metallo-
proteinase and increase tissue inhibitors of metallopro-
teinase expression in transformed cells. Rajagopalan et
al12 reported that incubation of cultured human vascular
smooth muscle cells with a superoxide-generating mix-
ture increases the amount of active matrix metallopro-
teinases, whereas nitric oxide donors did not have any
noticeable effect. In cultured cartilage, N-acetylcysteine
(10 mmol/L) and glutathione efficiently inhibit chon-
drolytic activity induced by fibronectin fragments.14 The
novelty of the current study is the extension of this con-
cept to amniochorionic membranes and its implication
for the pathogenesis of preterm premature rupture of
the membranes. Because generation of reactive oxygen
species by inflammatory cells is well documented20 and
infection is strongly associated with preterm labor, target-
ing formation of reactive oxygen species with thiol-reduc-
Buhimschi et al 463
ing agents may interfere with the pathophysiologic chain
that triggers preterm premature rupture of the mem-
branes and preterm birth.
Fetal membranes are composed of a single layer of
cuboidal amniotic epithelium of ectodermal origin, which
is in contact with the amniotic fluid at the apical pole. The
basal pole rests on a basement membrane and is separated
by a collagen- and fibroblast-rich layer from the chorion
(mesodermal derivative), which in turn is in intimate con-
tact with maternal decidua and trophoblasts.7 A significant
decrease in collagen content and an increase in collage-
nase activity occur close to term.7
Amniochorionic membranes from nonlaboring
women express only MMP-2.6 Culture conditions (even
of uninfected membranes) and intra-amniotic infections
induce MMP-9.6 Our results are consistent with those of
Fortunato et al6 because we examined matrix metallo-
proteinase levels after 24-hour culture in vitro. Our study
differs from that of Fortunato et al6 in that membranes
were incubated immediately after collection and only
after brief washing. In contrast, Fortunato et al6 ex-
changed the medium several times over a 48-hour period
before the additional 24-hour incubation in lipopolysac-
charide-containing media.21 We reasoned that an imme-
diate incubation would reflect more closely the in utero
conditions at term. However, under these conditions it is
possible that the membranes are actually maximally stim-
ulated (because of culture conditions immediately after
tissue collection) and cannot be further activated by
lipopolysaccharide. Another possibility is the simultane-
ous induction of both matrix metalloproteinases and tis-
464 Buhimschi et al
sue inhibitors of metalloproteinase by in vitro
lipopolysaccharide exposure, which was confirmed at
messenger ribonucleic acid levels with little6 or no
change (in our case) of matrix metalloproteinase levels,
as illustrated by substrate gel electrophoresis.
The results of our current study suggest that superox-
ide anion increases MMP-9 in human fetal membranes
and N-acetylcysteine inhibits matrix metalloproteinase ac-
tivity by preventing intracellular formation in the amnio-
chorion. However, it is possible that the effects of N-acetyl-
cysteine on matrix metalloproteinase activity extend
beyond inhibition of reactive oxygen speciesinduced ef-
fects, because treatment with N-acetylcysteine inactivates
MMP-2 in parallel with MMP-9. Moreover, the ability of N-
acetylcysteine to inhibit matrix metalloproteinase activity
becomes more valuable, as it involves total gelatinolytic
activity examined by zymography (the result of matrix
metalloproteinase release, matrix metalloproteinase acti-
vation, and matrix metalloproteinasetissue inhibitors of
metalloproteinase interaction).
It has been suggested that the mechanism by which re-
active oxygen species act as messenger molecules in
mammalian cells is through activation of transcription
factors, such as nuclear factor kb (NFkb) or activator
protein 1 (AP-1).22 Pretreatment of cells with antioxi-
dants such as N-acetylcysteine efficiently inhibits both cy-
tokine-induced NFkb activation cascade23 and induction
o f
AP-1.24 Both tyrosine kinasemediated NFkb and c-Jun
AP-1 activation are essential to the induction of MMP-9
by cytokines.25
In conclusion, amniochorionic matrix metallopro-
teinase activity is regulated by reduction-oxidation condi-
tions in the environment. MMP-9 levels in human fetal
membranes are directly increased by superoxide radical,
a byproduct of macrophages and neutrophils. Neither ni-
tric oxide donors nor nitric oxide synthase inhibitors sig-
nificantly affect matrix metalloproteinase activity in
human fetal membranes. The glutathione precursor N-
acetylcysteine dramatically inhibits amniochorionic ma-
trix metalloproteinase activity, in addition to inhibiting
intrinsic superoxide generation within the tissue. Thus
thiol-reducing agents, such as N-acetylcysteine, may be
beneficial in preventing preterm premature rupture of
the membranes associated with inflammation.
REFERENCES
1. Berkovitz GS, Papiernik E. Epidemiology of preterm birth.
Epidemiol Rev 1993;15:414-3.
2. Haywood JL, Goldenberg RL, Bronstein J, Nelson KG, Carlo
WA. Comparison of perceived and actual rates of survival and
freedom from handicap in premature infants. Am J Obstet
Gynecol 1994;171:432-9.
3. Romero R, Avila C, Brekus CA, Mazor M. The role of systemic in-
trauterine infection in preterm parturition. In: Garfield RE, ed-
itor. Uterine contractility. Norwell (MA): Serono Symposia;
1990. p. 319-54.
4. Oyarzun E, Yamamoto M, Kato S, Gomez R, Lizama L, Moenne
February 2000
Am J Obstet Gynecol
A. Specific detection of 16 micro-organisms in amniotic fluid by
polymerase chain reaction and its correlation with preterm de-
livery occurrence. Am J Obstet Gynecol 1998;179:1115-9.
5. Birkedal-Hansen H, Moore WG, Bodden MK, Windsor LJ,
Birkedal-Hansen B, DeCarlo A, et al. Matrix metalloproteinases:
a review. Crit Rev Oral Biol Med 1993;4:197-250.
6. Fortunato SJ, Menon R, Lombardi SJ. Collagenolytic enzymes
(gelatinases) and their inhibitors in human amniochorionic
membrane. Am J Obstet Gynecol 1997;177:731-41.
7. Parry S, Strauss JF 3d. Premature rupture of the fetal mem-
branes. N Engl J Med 1998;338:663-70.
8. Athayde N, Edwin SS, Romero R, Gomez R, Maymon E, Pacora
P, et al. A role for matrix metalloproteinase-9 in spontaneous
rupture of the fetal membranes. Am J Obstet Gynecol
1998;179:1248-53.
9. Osmers R, Tschesche H, Rath W, Szeverenyi M, Suwer V, Wolker
I, et al. Serum collagenase levels during pregnancy and parturi-
tion. Eur J Obstet Gynecol Reprod Biol 1994;53:55-7.
10. Esteve PO, Tremblay P, Houde M, St-Pierre Y, Mandeville R. In
vitro expression of MMP-2 and MMP-9 in glioma cells following
exposure to inflammatory mediators. Biochim Biophys Acta
1998;1403:85-96.
11. Gottschall PE, Yu X, Bing B. Increased production of gelatinase
B (matrix metalloproteinase-9) and interleukin-6 by activated
rat microglia in culture. J Neurosci Res 1995;42:335-42.
12. Rajagopalan S, Meng XP, Ramasamy S, Harrison DG, Galis ZS.
Reactive oxygen species produced by macrophage-derived foam
cells regulate the activity of vascular matrix metalloproteinases
in vitro. Implications for atherosclerotic plaque stability. J Clin
Invest 1996;98:2572-9.
13. Pardo A, Barrios R, Maldonado V, Melendez J, Perez J, Ruiz V, et
al. Gelatinases A and B are up-regulated in rat lungs by subacute
hyperoxia: pathogenetic implications. Am J Pathol 1998;
153:833-44.
14. Homandberg GA, Hui F, Wen C. Fibronectin fragment medi-
ated cartilage chondrolysis. I. Suppression by anti-oxidants.
Biochim Biophys Acta 1996;1317:134-42.
15. Rathakrishnan C, Tiku K, Raghavan A, Tiku ML. Release of oxy-
gen radicals by articular chondrocytes: a study of luminol-de-
pendent chemiluminescence and hydrogen peroxide secretion.
J Bone Miner Res 1992;7:1139-48.
16. McCord JM, Fridovich I. Superoxide dismutase. An enzymic
function forerythrocuprein (hemocuprein). J Biol Chem
1969;244:6049-55.
17. Telfer JF, Thomson AJ, Cameron IT, Greer IA, Norman JE.
Expression of superoxide dismutase and xanthine oxidase in my-
ometrium, fetal membranes and placenta during normal human
pregnancy and parturition. Hum Reprod 1997;12:2306-12.
18. Babior BM, Kipnes RS, Curnutte JT. Biological defense mecha-
nisms. The production by leukocytes of superoxide, a potential
bactericidal agent. J Clin Invest 1973;52:741-4.
19. Tyagi SC, Kumar S, Borders S. Reduction-oxidation (redox)
state regulation of extracellular matrix metalloproteinases and
tissue inhibitors in cardiac normal and transformed fibroblast
cells. J Cell Biochem 1996;61:139-51.
20. McCord JM, Wong K, Stokes SH, Petrone WF, English D.
Superoxide and inflammation: a mechanism for the anti-inflam-
matory activity of superoxide dismutase. Acta Physiol Scand
Suppl 1980;492:25-30.
21. Fortunato SJ, Menon R, Swan KF, Lyden TW. Organ culture of
amniochorionic membrane in vitro. Am J Reprod Immunol
1994;32:184-7.
22. Los M, Droge W, Stricker K, Baeuerle PA, Schulze-Osthoff K.
Hydrogen peroxide as a potent activator of T lymphocyte func-
tions. Eur J Immunol 1995;25:159-65.
23. Khachigian LM, Collins T, Fries JW. N-acetyl cysteine blocks
mesangial VCAM-1 and NF-kappa B expression in vivo. Am J
Pathol 1997;151:1225-9.
24. Bergelson S, Pinkus R, Daniel V. Intracellular glutathione levels
regulate Fos/Jun induction and activation of glutathione S-
transferase gene expression. Cancer Res 1994;54:36-40.
25. Yokoo T, Kitamura M. Dual regulation of IL-1 beta-mediated ma-
trix metalloproteinase-9 expression in mesangial cells by NF-