Human fetal membrane is composed of chorion, amnion, and interstitial tissues that maintain structural and mechanical integrity of the amniotic cavity during pregnancy. Failure of those biomechanical properties can result in rupture of the membrane and can influence labor outcome. Preterm premature rupture of the membranes (PROM) is associated with a high risk of neonatal morbidity and mortality.1 The exact cause of preterm PROM is unknown; however, many studies suggest intra-amniotic infection.2 We theorized that intra-amniotic infection engenders inflammatory responses that stimulate matrix metalloproteinase release and activation, causing weakening of the membranes and early rupture.
Draper and colleagues3 recently reported increased activity of proteases in fetal membranes collected from women who had PROM compared with membranes from women with normal gestations. In an earlier report we showed that human fetal membrane cells produce matrix metalloproteinases 2 and 9 (gelatinases). We documented induction of matrix metalloproteinase 9 during infection.4 Vadillo-Ortega and colleagues5 reported similar findings and induction of matrix metalloproteinase 9 in fetal membranes during labor. We also documented expression of tissue inhibitors of metalloproteinases (1, 2, 3, and 4) in human fetal membranes, suggesting a fully functional matrix metalloproteinase-tissue inhibitor of metalloproteinase network is active in human fetal membrane tissue.4,6
The stromelysin group of matrix metalloproteinases includes stromelysin 1 (matrix metalloproteinase 3), stromelysin 2 (matrix metalloproteinase 10), stromelysin 3 (matrix metalloproteinase 11), and matrilysin (matrix metalloproteinase 7). Those matrix metalloproteinases are capable of degrading many matrix components including type IV collagen, laminin, and proteoglycan. Stromelysin 1 cleaves procollagenase (matrix metalloproteinase 1) to its active form,7 and stromelysin 2 activates human neutrophil procollagenase (matrix metalloproteinase 8).8 Matrilysin has a broad and potent capacity to degrade numerous matrix components, including proteoglycans, insoluble elastin, and fibronectin, and is believed to be produced mainly by epithelial cells.9 The combined proteolytic activity of those matrix metalloproteinases (gelatinases and stromelysins) would be able to degrade essentially all the major matrix components of fetal membranes. Basement membrane collagen turnover is extremely important in maintaining integrity of the membrane, and it is likely that stromelysins and gelatinases are involved in the process.
Only stromelysin 1 has been reported in human fetal membranes.10 We report the presence of that group of matrix metalloproteinases in fetal membrane tissues under laboring and nonlaboring conditions, and in culture simulating intra-amniotic infections. We also report increased amniotic fluid (AF) levels of stromelysin 1 during preterm PROM compared with term gestations.
Materials and Methods
This study was approved by the institutional review board at The Women's Hospital at Centennial Medical Center, Nashville, Tennessee. Amniochorionic membranes were collected from women not in labor who had elective repeat cesareans (n = 8) and used for reverse transcriptase-polymerase chain reaction (PCR) screening and culture. Collection of amniochorionic membrane was described in detail previously.11 Those membranes were cultured in an organ explant system as described.11 Tissues were incubated for 48 hours then stimulated with bacterial lipopolysaccharide. In our earlier work with cytokines and matrix metalloproteinases, we documented that manipulating the membranes for culture induces expression of certain genes that are normally quiescent. It takes 48 hours in culture to bring the membrane back to a resting level of expression. Lipopolysaccharide (from Escherichia coli strain of O55: B5; Sigma, St. Louis, MO) at a concentration of 1000 ng/mL was used for stimulation. Tissues were harvested at the end of a 24-hour stimulation period, frozen in liquid nitrogen, and stored at −70C until analyzed for RNA. Media samples collected from each insert were stored separately in 1.5-mL tubes at −20C until they were assayed.
Additional membranes (n = 16) were collected from women after vaginal delivery and used only for reverse transcriptase PCR screening. After being dissected from placentas, membranes for reverse transcriptase PCR were cleansed of decidua and blood cells in Hank's balanced saline solution and frozen immediately in liquid nitrogen for RNA analysis. Histologic staining was done on sections collected from laboring women to rule out signs of histologic chorioamnionitis (over 3–4 polymorphonuclear leukocytes per high power field). All tissues used in this study were from women without histologic chorioamnionitis.
For basic reverse transcriptase PCR, RNA was extracted using TRIzol reagent (GIBCO-BRL, Gaithersburg, MD) according to manufacturer's instructions. A total of 0.5 μg of RNA was subjected to random primed (GIBCO-BRL) cDNA synthesis followed by 30 cycles of PCR in a Perkin-Elmer thermocycler (Perkin-Elmer Cetus, Norwalk, CT) using specific primers for stromelysin 1, stromelysin 2, stromelysin 3 and matrilysin. The PCR parameters were denaturation at 94C for 45 seconds, annealing at 55C for 45 seconds, and extension at 72C for 1 minute. For matrilysin, annealing was done at 51C for 45 seconds. The PCR products were analyzed on an ethidium bromide-stained 1.5% agarose gel. The specificity of primers designed to amplify each matrix metalloproteinases was tested by sequence analysis of the PCR products (Biotech-Core, Palo Alto, CA). The sequencing data obtained were subjected to a sequence identity search with Genbank (National Institutes of Health, Bethesda, MD) nucleotide database using the basic local alignment search tool program.
Amniochorion in culture at various time points (24, 48, and 72 hours) and after lipopolysaccharide stimulation was harvested, fixed in Histochoice (Amersco, Solon, OH) and processed in a tissue processor for in situ hybridization and immunohistochemistry as described.4,6,11
In situ hybridization was done using biotinylated oligonucleotides as described.4,6,11 The same antisense (3′) oligonucleotide used for PCR was used as the hybridization probe. The sense strand (5′) was used as a negative control for the specificity of hybridization.
Immunohistochemistry was done on paraffin-embedded sections to localize stromelysin 1. Immunohistochemical analysis was not done for other matrix metalloproteinases tested because specific antibodies are not commercially available. Mouse monoclonal antihuman stromelysin 1 antibody was used as the primary antibody (Cal Biochem, San Diego, CA). Sections were deparaffinized and rehydrated in graded alcohol and phosphate-buffered saline (pH 7.4). Rehydrated sections were quenched in 3% hydrogen peroxide in methanol to block endogenous peroxidase activity, and rinsed in phosphate-buffered saline, then exposed to 10% fetal bovine serum in phosphate-buffered saline for 10 minutes to minimize nonspecific antibody binding. Antigenic determinants masked by formalin fixation and paraffin embedding were exposed by enzymatic digestion. To achieve that, sections were treated with 0.1% trypsin (Cal Biochem, San Diego, CA) in phosphate-buffered saline (pH 7.4) for 12 minutes at room temperature, which was terminated by incubating the sections with 0.1 μg/mL of soybean trypsin inhibitor for 5 minutes followed by three washes in phosphate-buffered saline. Sections were incubated in ahumidified chamber with primary antibody (0.01 μg/mL) to stromerysin 1 for 4 hours. The manufacturer's instructions were followed for staining. Phosphate-buffered saline without primary antibody was used as a reagent control. Antibodies to HLA-DR and collagen type IV were used as negative and positive controls, respectively.
As a preliminary clinical correlation to our in vitro work, a cross-sectional study was designed to evaluate the relationship between PROM and stromelysin 1 levels in AF. Amniotic fluid samples from eligible women were selected from our bank of previously collected and frozen samples. That study included women with preterm PROM (n = 15), which was defined as spontaneous amniorrhexis before labor and before 37 weeks' gestation. Membrane rupture was diagnosed using positive identifiers such as nitrazine test, fern test, pooling, and oligohydramnios on ultrasound examination. The second group (n = 25) consisted of women at term who had elective repeat cesareans with no history of pregnancy complications. Women with other medical or obstetric complications and those with fetuses with anomalies were excluded.
All AF samples were collected by transabdominal amniocentesis under continuous ultrasonographic guidance or through intact fetal membranes before uterine incisions during cesareans. Samples were centrifuged at 1700 × g for 10 minutes and the supernatants were stored at −70C until ready for assay.
Stromelysin 1 levels in culture media and AF samples were quantitated using enzyme-linked immunosorbant assay (ELISA). The levels of other matrix metalloproteinases in this study were not quantitated because reagents were not commercially available at the time. The assay procedure involved a multiple-site, two-step sandwich immunoassay using monoclonal antibody (Amersham Life Sciences, Buckinghamshire, England). The manufacturer's instructions were followed for ELISA. Standard curves were developed using duplicate samples of known quantities of recombinant human stromelysin 1. Concentrations were determined by relating absorbance to the standard curve by linear regression analysis. Controls consisted of assay buffer, plain media, and lipopolysaccharide-containing media incubated in wells without tissue. Colorimetric absorption was read at 405 nm using an IncStar Automatic microplate reader (IncStar Corporation, Still Water, MN). The assay detects immunoreactive and bioreactive forms of stromelysin 1.
Statistical significance of the data from ELISA on culture media samples and AF samples were analyzed by unpaired, nonparametric, two-sample Mann-Whitney U test. A two-tailed P ≤ .05 was considered statistically significant.
Messenger RNA for stromelysin 1, stromelysin 2, and stromelysin 3 was detected in all tissues tested in this study. Reverse transcriptase PCR data showed that amniochorionic membranes collected from women in labor at term and women at term who had cesareans expressed mRNA for all three stromelysins (Figure 1). When membranes from women not in labor were placed in culture, stromelysin expression was evident up to 3 days (the longest tissues were cultured in the study) (Figure 2). Lipopolysaccharidestimulated tissues did not show any visible changes in the expression pattern of stromelysins (data not shown) by basic PCR; however, quantitation of mRNA was not done on those samples. Matrilysin mRNA was not expressed in tissues from women in labor (n = 16) or women not in labor (n = 8). Neither in vitro culture conditions nor stimulation with various doses of lipopolysaccharide (n = 8) was associated with matrilysin expression (data not shown). The matrilysin primer specificity was confirmed using RNA from a breast carcinoma cell line (MDA-MB-468) known to produce matrilysin mRNA.
The specificity of primers used in those reverse transcriptase PCR experiments was confirmed by sequencing of PCR products. The data were subjected to sequence identity search using the Genbank database (National Institutes of Health, Bethesda, MD) and found to amplify the specific gene of interest (data not shown). DNA contamination and amplification of genomic DNA was ruled out by cDNA synthesis, without reverse transcriptase enzyme, followed by PCR.
Messenger RNA for stromelysins 1, 2, and 3 was identified in amnion and chorion cells (both laeve and cytotrophoblast) in all tissues tested. Cells in the spongy layer and compact layer of the extracellular matrix showed hybridization with antisense strand to the respective mRNAs. Hybridization in the extracellular matrix region was scattered in all the tissues analyzed. The sense strand was used as a negative control for the specificity of the reaction. No nonspecific hybridization for stromelysin 1 or stromelysin 2 was seen; however, a slight amount of nonspecific hybridization was seen in chorion when the sense strand for stromelysin 3 was used (Figure 3).
Immunohistochemical localization was done with monoclonal antibody to stromelysin 1. Immunohistochemical analysis localized the stromelysin 1 to amnion and chorion (laeve and cytotrophoblast) cells in the spongy and reticular layers of the membranes (Figure 4). Those data agreed with the in situ hybridization data. Immunohistochemical data using the positive and negative control antibodies was as expected. Antibodies to collagen type IV (positive control) localized the presence of that protein to the chorion pseudobasement membrane and also to chorionic cytotrophoblast cells. No staining was seen in amnion basement membrane. Phosphate-buffered saline and antibody to HLA-DR (negative controls) documented no nonspecific staining reactions. Data were consistent for all tissue preparations tested.
In vitro, lipopolysaccharide stimulation increased stromelysin 1 release from fetal membranes compared with control. Lipopolysaccharide stimulation (1000 ng/mL), significantly increased stromelysin levels (median 70.35 ng/mL, range 13.42–112 ng/mL) compared with control (median 15.8 ng/mL, range 1.45–69 ng/mL) (P = .05). Even low-level lipopolysaccharide stimulation (50 ng/mL) nearly achieved significance (P = .06).
A total of 40 AF samples were collected for stromelysin 1 analysis. The mean gestational age for the preterm PROM group was 31 weeks (standard deviation [SD] ± 3.2 weeks) and for the term group was 37.72 weeks (SD ± 1.12 weeks) (P = .001). In the preterm PROM group, microbiologic evaluation was done. Three of those samples were culture-positive for microorganisms (Group B β hemolytic streptococci-2 Klebsiella oxytoca-1). The stromelysin levels for the culture-positive subgroups did not differ significantly from the overall group (P = .5). Preterm PROM was associated with an increase (median 3.2 ng/mL, range 0–102.9 ng/mL) in the levels of stromelysin 1 compared with amniotic fluid from women at term (median 1.3 ng/mL, range 0–26.8 ng/mL) (Mann-Whitney P = .02).
The expression of all three stromelysins, the gelatinases, and all of the known tissue inhibitors of metalloproteinases has been documented in human fetal membrane tissue.4,6,10,11,12 It was proposed originally that bacterial proteases produced during intra-amniotic infection were responsible for PROM.13 However, in vivo species and strain-specific enzyme contributions in PROM are still unknown. In view of the present data and our earlier reports on collagenolytic enzymes and tissue inhibitor of metalloproteinases in the membrane, we propose that inflammatory response to intra-amniotic infection might contribute to activation of endogenous matrix-degrading enzymes, which perturbs the matrix metalloproteinase-tissue inhibitor of matrix metalloproteinase balance, leading to increased matrix metalloproteinase activity and preterm PROM, as is supported by our in vitro findings (unpublished data). In vitro amniochorion can respond to infectious stimuli by releasing immunoactive and bioactive stromelysin 1 and gelatinases, all of which can degrade extracellular matrix components. The tissue inhibitors of matrix metalloproteinases decrease in AF of women with PROM (unpublished data). In vivo, a significant change in stromelysin levels was not seen during intra-amniotic infection, although that might be due to small sample size. Only three amniotic fluids were culture positive, and it is difficult to comment on the exact effect of bacteria on inducing stromelysin in vivo. We found that concentrations of stromelysin 1 produced by the small tissue circles in culture were significantly higher than the concentrations seen in amniotic fluids, which might indicate that amniotic fluid levels are a poor indicator of what is occurring at the tissue level.
Stromelysins might function in several ways to achieve matrix turnover or degradation. They can cleave collagen type IV in the basement membrane, which connects amnion and chorion to the extracellular matrix. In our immunohistochemical analysis, type IV collagen was in the basement membrane and in the spongy and reticular layers of the extracellular matrix (positive controls in our immunohistochemical studies, data not shown). Proteoglycan, laminin, and elastin, all of which add to the structural stability of those regions, are also degraded by the stromelysins. Degradation of those matrix components and eventual recycling or destruction of basement membrane is most likely a synergistic activity of the stromelysins in concert with the gelatinases. Fibrillar collagen, seen in the extracellular matrix, can be broken down by interstitial collagenase (matrix metalloproteinase 1), which requires proteolysis by stromelysin 1 for activation. Ogata et al14 reported that stromelysin 1 can also cleave inactive 92-kDa gelatinase (matrix metalloproteinase 9) into its active 82-kDa form, which might be significant during infection and labor when fetal membranes induce expression of 92-kDa gelatinase.
Histologic chorioamnionitis (defined as polymorphonuclear leukocyte invasion of the membranes) is a common finding in preterm PROM associated with microbial invasion of the amniotic cavity. Activation of neutrophil collagenase in those polymorphonuclear leukocytes is an effect of stromelysin 2. Stromelysin 3 has not been associated with any substrate-specific degradation to date.15 Unlike the other two stromelysins, it is not known whether stromelysin 3 can directly degrade any extracellular matrix substrate. Similar to other matrix metalloproteinases, stromelysin 3 can cleave α1-antitrypsin, indicating that it has proteolytic activity.16
PROM is a serious problem because it is difficult to predict and its causes are still unknown.1,2 During PROM, increased matrix turnover might lead to membrane destruction. Our recent findings indicate that there is an overwhelming presence of gelatinases (matrix metalloproteinase 2 and matrix metalloproteinase 9) compared with their counter-regulatory proteins during an infectious process,14 which suggests that increased stromelysins from fetal membranes with gelatinases might affect membrane degradation during PROM.
1. Gibbs RS, Blanco JD. Premature rupture of the membranes. Obstet Gynecol 1982;60:671–9.
2. Gomez R, Romero R, Edwin SS, David C. Pathogenesis of preterm labor and preterm premature rupture of membranes associated with intraamniotic infection. Infect Dis Clin North Am 1997;11:135–76.
3. Draper D, McGregor J, Hall J, Jones W, Beutz M, Heine RP, et al. Elevated protease activities in human amnion and chorion correlate with preterm premature rupture of membranes. Am J Obstet Gynecol 1995;173:1506–12.
4. Fortunato SJ, Menon R, Lombardi SJ. Collagenolytic enzymes (gelatinases) and their inhibitors in human amniochorionic membrane. Am J Obstet Gynecol 1997;177:731–41.
5. Vadillo-Ortega F, Gonzalez-Avila G, Furth EE, Lei H, Muschel RJ, Stetler-Stevenson WG, et al. 92-kd type IV collagenase (matrix metalloproteinase-9) activity in human amniochorion increases with labor. Am J Pathol 1995;146:148–56.
6. Fortunato SJ, Menon R, Lombardi SJ. Presence of four tissue inhibitors of matrix metalloproteinases (TIMP-1, -2, -3 and -4) in human fetal membranes. Am J Reprod Immunol 1998;40:395–400.
7. Murphy G, Cockett MI, Stephens PE, Smith BJ, Docherty AJ. Stromelysin is an activator of procollagenase. A study with natural and recombinant enzymes. Biochem J 1987;248:265–8.
8. Knauper V, Murphy G, Tschesche H. Activation of human neutrophil procollagenase by Stromelysin 2. Eur J Biochem 1996;235:187–91.
9. Wilson CL, Matrisian LM. Matrilysin: An epithelial matrix metalloproteinase with potentially novel functions. Int J Biochem Cell Biol 1996;28:123–36.
10. Bryant-Greenwood GD, Yamamoto SY. Control of peripartal collagenolysis in the human chorion-decidua. Am J Obstet Gynecol 1995;172:63–70.
11. Fortunato SJ, Menon R, Swan KF, Lyden TW. Organ culture of amniochorionic membrane in vitro. Am J Reprod Immunol 1994; 32:184–7.
12. 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.
13. McGregor JA, French JI, Lawellin D, Franco-Buff A, Smith C, Todd JK. Bacterial protease-induced reduction of chorioamniotic membrane strength and elasticity. Obstet Gynecol 1987;69:167–74.
14. Ogata Y, Enghild JJ, Nagase H. Matrix metalloproteinase 3 (stromelysin) activates the precursor for the human matrix metalloproteinase 9. J Biol Chem 1992;267:3581–4.
15. Basset P, Bellocq JP, Wolf C, Stoll I, Hutin P, Limacher JM, et al. A novel metalloproteinase gene specifically expressed in stromal cells of breast carcinomas. Nature 1990;348:699–704.
16. Pei D, Majumdar G, Weiss SJ. Hydrolytic inactivation of a breast carcinoma cell-derived serpin by human stromelysin-3. J Biol Chem 1994;269:25849–55.