- Email: [email protected]
Distinct molecular events suggest different pathways for preterm labor and premature rupture of membranes
Distinct molecular events suggest different pathways for preterm labor and premature rupture of membranes Stephen J. Fortunato, MD, and Ramkumar Menon, MS Nashville, Tennessee OBJECTIVE: On a clinical level, the etiologies associated with premature rupture of the membranes and preterm labor are virtually identical, though these conditions end in distinctly different events. This study was designed to determine differences between preterm labor and preterm premature rupture of membranes by using molecular markers of extracellular matrix degradation and apoptosis. STUDY DESIGN: Amniochorion and amniotic fluid samples were collected from gestational age–matched groups of women undergoing cesarean delivery before term. Samples were collected from 2 groups of women, women with premature rupture of membranes and women with preterm labor with no rupture of membranes. Changes in the expression pattern of messenger ribonucleic acid for matrix metalloproteinases (MMP), tissue inhibitor of metalloproteinases (TIMP), and pro-apoptotic (p53 and Bax) and anti-apoptotic (Bcl-2) proteins were identified by quantitative polymerase chain reaction. Enzyme-linked immunosorbent assay was used to determine the levels of these proteins in the amniotic fluid. Multiplex polymerase chain reaction was performed to study the expression of Fas-Fas ligand–associated pro-apoptotic genes. Unpaired nonparametric, 2-tailed Mann-Whitney U test was used to determine statistical significance of quantitative polymerase chain reaction and enzyme-linked immunosorbent assay (P < .05 was considered significant). RESULTS: Quantitative polymerase chain reaction results demonstrated an increased mRNA expression for MMP2, MMP9, and MT1-MMP and a decreased expression for TIMP2 in prematurely ruptured membranes compared with preterm labor membranes. Enzyme-linked immunosorbent assay documented increases in the amniotic fluid concentrations of immunoreactive and bioactive MMP2 and MMP9 and immunoreactive MMP3 and a decreased TIMP2 concentration in fluids obtained from the premature rupture of membranes group compared with the preterm labor group. The pro-apoptotic genes p53 and bax were up-regulated in premature rupture of membranes when compared with preterm labor. Anti-apoptotic gene (Bcl-2) expression was increased in preterm labor membranes compared with prematurely ruptured membranes. Interleukin-18 (a pro-apoptotic cytokine) was increased in the amniotic fluid during premature rupture of membranes compared with preterm labor. Prematurely ruptured membranes also demonstrated fragmented deoxyribonucleic acid and expression of Fas and caspase 8 (apoptosis initiator), which were all absent in preterm labor membranes. CONCLUSIONS: We have begun to delineate 2 divergent molecular pathways for premature rupture of membranes and preterm labor. Most likely, this is the beginning of the identification of differences that will become evident with the use of molecular biology. (Am J Obstet Gynecol 184:1399-406.)
Key words: Premature rupture of membranes, fetal membranes, apoptosis, preterm labor, cytokines, matrix metalloproteinases
Preterm birth complicates 10% to 15% of all pregnancies, and it accounts for high fraction of the perinatal morbidity and mortality rate.1 Spontaneous preterm From The Perinatal Research Center of The Women’s Health Research and Education Foundation, The Women’s Hospital at Centennial Medical Center, Nashville, Tennessee. Supported in part by grants from Columbia/HCA Healthcare Foundation. Presented at the Sixty-eighth Annual Meeting of The Central Association of Obstetricians and Gynecologists, Chicago, Illinois, October 18-21, 2000. Reprint requests: Stephen J. Fortunato, MD, The Perinatal Research Center of the Women’s Health Research and Education Foundation, 2300 Patterson St, Nashville, TN 37203. Copyright © 2001 by Mosby, Inc. 0002-9378/2001 $35.00 + 0 6/6/115122 doi:10.1067mob.2001.115122
birth, caused by preterm labor (PTL) (contractions before 36 weeks’ gestation) or preterm premature rupture of the membranes (PROM) (membrane rupture before the onset of labor) or both account for ~80% of preterm deliveries.1 PROM is associated with 50% of the PTL cases. The frequency of premature birth has actually increased over the past 3 decades, which in part can be attributed to the fact that the exact cause of these conditions is still unknown. The etiology of PTL and PROM is multifactorial, and most of the causes are still unknown. Microbial invasion of the intra-amniotic cavity and infection of the intraamniotic cavity are associated with a majority of these cases. Infection induces a maternal and fetal inflammatory response (histologic chorioamnionitis), increases 1399
1400 Fortunato and Menon
production and release of the inflammatory cytokines (interleukin [IL]-1, IL-6, IL-8, and tumor necrosis factor-α [TNF-α]), which induce prostaglandin production leading to cervical ripening and prostaglandin-mediated contractility. Elevations of these biologic messengers (cytokines and prostaglandins) are considered markers of PTL and PROM. These factors are common to both PTL and PROM and are elevated in the amniotic fluid even in the absence of microbial invasion of the intra-amniotic cavity and infection of the intra-amniotic cavity when compared with term labor.2-8 The puzzling question we address here is why some women experience PROM and some experience PTL with no rupture of membranes when the etiologic factors associated with both these pathologic complications are the same. Infection, smoking, socioeconomic factors, racial differences, nutritional status, multiple gestation, polyhydramnios, and other conditions have all been associated with both PTL and PROM. To date, studies had evaluated the markers that are commonly elevated in both PTL and PROM. A better understanding of the similarities and differences between the pathways leading to each of these conditions may open new avenues for research and treatment. It has been known for a long time that the inflammatory cytokines are elevated in both PTL and PROM and are most likely instrumental in promoting uterine contractility through increased prostaglandin release. More recently, studies from Draper et al,9 Vadillo-Ortega et al,10, 11 Maymon et al,12 and our laboratory13 have shown that intrinsic activation of matrix metalloproteinases (MMPs) is involved in PROM.14, 15 The presence or absence of apoptosis also appears to represent a significant difference between these processes.15, 16 In this article we review differences in the molecular events associated with PROM and PTL. Methods Amniochorionic membranes were collected from the following groups of women at the time of cesarean delivery for obstetric indications. Group 1 consisted of women with PROM who were delivered by cesarean before term (n = 10). Group 2 consisted of women with PTL and no rupture of membranes (n = 10) who were delivered by cesarean before term. A cross-sectional study was designed to evaluate the changes in the levels of MMPs and cytokines in the amniotic fluid during PROM and PTL. Amniotic fluid samples from eligible women were selected from our bank of previously collected and frozen samples. This study included women with preterm PROM, which was defined as spontaneous amniorrhexis before labor and before 37 weeks’ gestation. A positive Nitrazine test result, a positive fern test result, or positive pooling and oligohydramnios on ultrasound examinations diagnosed membrane rupture. The second group consisted
June 2001 Am J Obstet Gynecol
of women with PTL and no rupture of membranes before 37 weeks’ gestation. Women with other medical or obstetric complications were excluded from this study. All amniotic fluid samples were collected by transabdominal amniocentesis under continuous ultrasonographic guidance or through uterine incisions at the time of cesarean delivery. Samples were centrifuged at 1700 rpm for 10 minutes, and the supernatants were stored at –70ºC until analysis. Amniotic fluid samples were also subjected to culture for microorganisms. Only a very few samples were positive for microorganisms (3 in each group: group B streptococci, 3, Klebsiella oxytoca, 1; Staphylococcus aureus, 1; and Candida species, 1). Accordingly, in this report culturepositive and culture-negative sample groups were combined. Histologic examination of the fetal membranes was performed to document histologic chorioamnionitis (>3-4 neutrophils/high-power field), and the positive membranes were excluded from this study. Ribonucleic acid (RNA) extraction and complementary deoxyribonuecleic acid (DNA) synthesis. Total RNA was extracted by the Trizol (Life Technologies, Bethesda, Md) method, and 0.5 µg of total RNA was subjected to oligo dT primed reverse-transcriptase reaction. Complementary deoxyriblonucleic acid (DNA) was then subjected to quantitative polymerase chain reaction (QPCR) or multiple PCR with target-specific primers. QPCR for MMP2, MMP9, MT1-MMP, TIMP2, p53, bax, and Bcl-2. In the competitive PCR assay, one set of primers is used to amplify both the target gene complementary DNA and another neutral DNA fragment. A known concentration of neutral DNA fragment competes with the target cDNA fragment (unknown) for the same primers (primer sequences used in this study were published earlier) and acts as an internal standard. Detailed protocols can be found elsewhere.14, 16 Gene expression of MMPs, TIMP2, pro-apoptotic genes (p53 and bax) and anti-apoptotic (Bcl-2) factors were analyzed in PROM and PTL membranes by QPCR. A constant amount of cDNA was subjected to QPCR. Quantitation of QPCR gels was done by using densitometric analysis with the Alpha-Ease (AlphaInnotech Corp, San Leandro, Calif) software program.14,17 Enzyme-linked immunosorbent assay (ELISA) for MMPs 2, 3, and 9, TIMP2, and interleukin-18. The levels of MMPs (immunoreactive MMP2, 9 [gelatinases], MMP3 [stromelysin 1], and TIMP2) and the pro-apoptotic cytokine, IL-18, were quantitated in the amniotic fluid by means of ELISA. The assay procedure involved a multiple-site 2-step sandwich immunoassay with oligoclonal antibodies. These assays were performed in our laboratory with commercially available kits (Amersham-Pharmacia, Piscataway, NJ [for MMP 2, 3, 9 and TIMP2] R&D Systems, Minneapolis, Minn [for IL-18]). The manufacturer’s instructions were followed for each kit to perform
Fortunato and Menon 1401
Volume 184, Number 7 Am J Obstet Gynecol
Table I. Quantitative expression of MMPs in PROM and PTL by QPCR PROM (n = 8)
MMP2 MMP9 MT1-MMP TIMP2
PTL (n = 8)
Median
±SD
Range
Median
±SD
Range
P value
60,000 33 600 550,000
29,88 0.3 196 287,838
6000-600,000 0.6-60 65-662 55,000-660,000
600 0.6 0.6 535,000
2548 2.9 36.5 252,523
600-6000 0.6-6 0.6-80 50,000-640,000
.001 .02 .006 .9
Data are presented as number of transcripts per 0.5 µg of total RNA.
ELISA. Standard curves were developed by using duplicate samples of known quantities of recombinant proteins. Sample concentrations were determined by relating the absorbance obtained to the standard curve by linear regression analysis. Colorimetric absorption was read at 450 nm by using a microplate reader. Controls consisted of assay buffer. The sensitivities of various assays were as follows: MMP2, 0.37 ng/mL; MMP3, 1.175 ng/mL; MMP9, 0.2 ng/mL; and TIMP2, 1.5 ng/mL. Active MMP9 in amniotic fluid. The amniotic fluid concentration of inhibitor-free, active MMP9 was tested by using previously described methods.15 A novel protein, engineered to contain an MMP9-specific cleavage site, was used as substrate in this assay (Amersham-Pharmacia). The cleavage products were colorimetrically quantitated in an ELISA format after capture onto the wells by incubating at 4ºC overnight in anti-MMP9 antibody– coated microtiter plates. After washing, the plates were incubated at 37ºC with the substrate protein and a detection enzyme for 6 hours. The resulting color was read at 405 nm by using an ELISA plate reader. Quantitation of zymograms for active MMP2. Amniotic fluid samples were subjected to gelatin zymography along with human recombinant active MMP2 standard (Oncogene Research Products, Cambridge, Mass). Different standard dilutions (range, 1 ng–0.0001 ng) were made and subjected to zymography. The quantification of the bands produced by these standards of known concentrations was used (for standard curve) to calculate the concentrations of active MMP2 (a 66-kd band) in amniotic fluid samples with the use of the Alpha-Ease software program.15 Multiplex PCR assay for Fas-FasL apoptotic pathway genes. Multiplex PCR (MPCR) was used to study the expression pattern of pro-apoptotic genes such as Fas, Fas ligand, caspase 8, and Fas-associated death domain in PROM and PTL amniochorionic membranes. In multiple PCR, several target genes are amplified simultaneously in the same reaction tube (Cytoexpress, Multiplex PCR, BioSource International, Camarillo, Calif). Multiplex primers are used to amplify different target transcripts at the same time. In this assay, Fas, Fas ligand, caspase 8, TNF receptor–associated death domain, and Fas-associated death domain gene expression was studied. Glyceralde-
hyde 3-phosphate dehydrogenase was used as the control gene. Complementary DNA from reverse transcriptase reactions was subjected to MPCR. Manufacturer’s instructions were followed for MPCR. DNA fragmentation analysis by ligand-mediated PCR. The extent of DNA damage during PROM and PTL caused by apoptosis of the cells was investigated by using the ligation-mediated PCR assay. The details of this assay protocol are described elsewhere.17 Statistical analysis. Unpaired nonparametric, 2-sample Mann-Whitney U test was used to analyze statistical significance of QPCR and ELISA (P < .05 was considered significant). Results The mean gestational age of fetuses in the PROM group was 31.6 ± 3.4 weeks and of the PTL group was 30 ± 4.3 weeks (P = 0.6). Histologic chorioamnionitis was documented in 40% (4/10) of prematurely ruptured membranes and 50% (5/10) of PTL membranes. The presence of histologic chorioamnionitis was not associated with a statistically significant difference in any of the factors studied in this report. Increased expression of MMPs in PROM. Amniochorionic membranes showed an increased expression of gelatinases (MMP2 and MMP9) and the MMP2 activator protein gene MT1-MMP (MMP14) in membranes from the PROM group compared with those from the PTL group (Table I). In prematurely ruptured membranes, 2.4 × 105 transcripts of MMP2 were seen, whereas only 1.8 × 103 transcripts were seen in PTL membranes. MT1MMP transcripts were similarly increased in prematurely ruptured membranes compared with PTL membranes (545.6 vs 20.9 transcripts, respectively; P = 0.006). MMP9 transcripts were increased from 2.4 in PTL membranes to 31.6 in prematurely ruptured membranes (P = .02). The MMP inhibitor TIMP2 messenger RNA levels were similar in PROM and PTL. The number of transcripts of TIMP2 was 3.5 × 105 in PROM compared with 4.2 × 105 in PTL (P = not significant). MMP and TIMP2 concentrations in amniotic fluid. Immunoreactive MMP9 concentration was significantly increased in the amniotic fluid samples obtained from women with PROM (PTL, 3.3 ng/mL; PROM, 15.25
1402 Fortunato and Menon
June 2001 Am J Obstet Gynecol
Table II. Amniotic fluid levels of MMPs and TIMP2 (n = 18) PROM
MMP2 (ng/mL) MMP9 (ng/mL) MMP 3 (ng/mL) A-MMP2* (pg/mL) A-MMP9* (pg/mL) TIMP2 (ng/mL) IL-18* (pg/mL)
PTL
Median
±SD
Range
Median
±SD
Range
P value (PROM vs PTL)
2400 8.6 1.5 200.5 200 85 1990
420 14.3 29.4 161.7 1130 43.7 3465.7
1300-2500 0-30.9 0-102.9 122-851 0-4900 60-250 710-1522
2050 0 1.1 105.5 0 184 888
879.8 9.5 17.16 85.0 200 130.4 693.4
170-2400 0-9.7 0-69.8 0-263 0-1000 100-560 170-2746
.1 .002 .8 .01 .04 .0001 .001
A, Active levels.
Table III. Quantitative expression of apoptotic factors during PROM and PTL PROM (n = 10)
PTL (n = 10)
Median
±SD
Range
Median
±SD
Range
P value
60 486.5 2.8
25.1 230.4 2.5
6-86 56-892 0-6
3 32 58
17.9 200.3 287.69
0.6-60 0.5-509 0-656
.001 .003 .05
p53 Bax Bcl-2
Data are presented as number of transcripts per 0.5 µg of total RNA.
Table IV. Expression of Fas-FasL–mediated apoptotic pathway genes during PROM and PTL
PROM PTL
Fas
Fas L
Caspase 8
FADD
TRADD
GAPDH
+ ±
– –
+ –
+ +
+ ±
+ +
FADD, Fas-associated death domain; TRADD, TNF receptor– associated death domain; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; ±, expression seen in at least 50% of the tissues.
ng/mL; P = 0.002). In fact, immunoreactive MMP9 was present in detectable concentrations in only 27% (5/18) of the PTL samples, whereas 72% (13/18) of the PROM samples contained detectable immunoreactive MMP9. Although both MMP2 and MMP3 amniotic fluid concentrations were increased in PROM, this was not a statistically significant finding, in part because of the large standard deviation associated with each group (Table II). A significant and inverse relationship was seen in TIMP2 concentrations when compared with the MMPs. The concentration of this MMP inhibitor was decreased in PROM (97.2 ng/mL) when compared with PTL (248.9 ng/mL; P = 0.0001). The molar ratios between MMP2 and its inhibitor/activator TIMP2 were also calculated. The ratio of MMP2 to TIMP2 was 2 in the amniotic fluid of the PTL group compared with 6.3 in the PROM group. TIMP2 was chosen for investigation because it performs the role of MMP2 activator at low equimolar concentrations and the role of MMP2 inhibitor at higher concentrations. The active, TIMP-free forms of MMP2 and 9 in amniotic fluid were assayed. Less than 1% of the total proteins were active in both PROM and PTL, and the active forms
were elevated in PROM compared with PTL. Zymogram quantitation of MMP2 revealed an active MMP2 level of 233.5 pg/mL in the amniotic fluid samples from PROM, whereas the mean level was 127.8 pg/mL in PTL (P = .01). Amniotic fluid active MMP9 levels were 600 pg/mL in PROM samples, whereas the levels were below the detection limits of the assay in most (83%) of the PTL samples (P = .04). Of note, Athayde et al18, 19 have reported active MMP9 in the amniotic fluid of women with PTL with intact membranes using an assay modified to detect lower concentrations. Their data are in agreement with our findings that active MMP9 levels are higher in PROM compared with PTL.12 A specific activity assay for MMP3 was not performed because such a method is not yet commercially available. Elevation of apoptotic elements during PROM. It is hypothesized that apoptosis may be responsible for the MMP increase in PROM because some pro-apoptotic proteins such as p53 are capable of transactivating some of the MMP genes.17 DNA fragmentation (intra-nucleosomal cleavage and laddering of DNA) (data not shown) was noticed in the majority of the prematurely ruptured membranes (8/10) compared with 30% (3/10) of PTL membranes.17 We quantitated the proapoptotic protein p53 induced by DNA damage, as well as p53 target proteins such as bax (pro-apoptotic) and Bcl-2 (anti-apoptotic), using QPCR (Table III). Compared with PTL, the expression of pro-apoptotic genes was elevated in the prematurely ruptured membranes. p53 transcripts were increased from 9.3 in PTL membranes to 49.6 (P = .003) in prematurely ruptured membranes, and bax transcripts were 120.8 and 524.3 in PTL and PROM, respectively (P = .004). The number of anti-
Volume 184, Number 7 Am J Obstet Gynecol
Fortunato and Menon 1403
Fig 1. Factors associated with PROM can increase expression of MT1-MMP, MMP2, and MMP3 from human fetal membranes and induce MMP9 expression or release. These factors reduce MMP inhibitor TIMP2. MT1-MMP and low levels of TIMP2 activate pro-MMP2 to active form. Active MMP2 and 3, along with various other proteases, activate MMP9. All these active MMPs can degrade extracellular matrix proteins to cause membrane weakness and rupture. Factors of MMP activation can also activate the apoptotic pathway mediated by p53 and TNF (see Fig 2). p53 is a known inducer of the MMP2 gene. Synergistic action of all these factors may activate a vicious cycle of events resulting in membrane rupture.
apoptotic protein Bcl2 gene transcripts was 100-fold higher in PTL compared with that of PROM (2.8; P = .05), suggesting predominance of an anti-apoptotic effect in PTL. Existence of Fas- and TNF-mediated apoptotic pathway. MPCR showed the expression of TNF- and Fas-mediated apoptotic pathway genes in PROM membranes (Table IV). Fas, an apoptosis inducer, was seen in the majority of prematurely ruptured membranes (9/10) and some PTL membranes (5/10). Caspase 8, an activator/initiator of the caspase cascade of cell death proteins, was expressed in a majority of the prematurely ruptured membranes (9/10), whereas only 30% (3/10) of PTL membranes showed caspase 8 gene expression. The docking proteins of the Fas- and TNF-mediated apoptotic signaling pathway (Fas-associated death domain; TNF-associated death domain) were present in many of the PTL (5 of 10) and prematurely ruptured (7 of 10) membranes. Fas ligand expression was never seen in fetal membranes during either condition. Glyceraldehyde 3-phosphate dehydrogenase was used as a “housekeeping gene” in this experiment and was present in all the membranes tested. The level of the pro-apoptotic and pro-inflammatory cytokine, IL-18, was also elevated in the amniotic fluid during PROM. The mean levels were 1067 pg/mL in PTL compared with 3230 pg/mL in PROM (P = .001). Comment In analyzing the data presented herein, it is obvious that distinct pathways leading to each of the conditions
studied (PROM and PTL) exist. Although it would seem obvious that this should be the case, there is a paucity of data addressing this area. The existence of this separation is made less obvious by the fact that each of these conditions has exactly the same set of etiologic factors. Although PROM and PTL frequently accompany each other, it is well known that they can present as totally separate entities. Both pathways are known to share activation of the cytokine pathway to a greater or lesser degree, it seems, depending on the presence or absence of infection. We have been able to identify several branch points that differentiate one from the other. Initially, our laboratory began looking at the role of the MMPs in PROM. It is now clear that the fetal membranes are endowed with the full system of MMPs, MMP inhibitors (TIMPs), MMP activators, and regulatory elements. Most of the MMPs are increased in membranes that have sustained premature rupture. The active forms of these MMPs are increased in PROM when compared with any other conditions in pregnancy. Several patterns of data lead us to speculate that active MMP9 (gelatinase B) plays a crucial role in PROM: (1) the MMP9 gene is induced in amniochorion during labor, PROM, and infection; (2) MMP9 levels are increased in the amniotic fluid as gestation progresses and also during active labor at term and before term; (3) gelatinolytic activity, corresponding to latent and active forms of MMP9, is increased in amniotic fluid obtained from patients with PROM; (4) a significant in-
1404 Fortunato and Menon
June 2001 Am J Obstet Gynecol
Fig 2. Two major apoptotic pathways may exist in PROM, which can be initiated by infection, genotoxic agents, or unknown factors: (1) TNF receptor–Fas–mediated pathway—these receptor proteins bind to their respective ligands TNF and Fas L, which initiate signal transduction through 2 docking proteins TRADD (TNF receptor– associated death domain) and FADD (Fas-associated death domain). These death domain proteins activate procaspase 8 to active caspase 8. (2) p53-mediated pathway is initiated by DNA fragmentation. DNA damage increases transactivator protein p53 in the cell. p53 transactivates bax, which causes damage to the mitochondrial membrane, resulting in the release of cytochrome c. Cytochrome c activates Apaf (apoptosis protease activating factor), which converts pro-caspase 9 to active form. p53 also suppresses Bcl-2, a factor that stops mitochondrial membrane damage. Active caspase 8 or 9 can initiate a cascade of caspase activation. Caspases 3, 7, and 6 are activated sequentially, which will cause proteolysis of structural proteins, proteins of homeostasis, and several other proteins and program the cell to death. These pathways cross over at several points. p53 can turn on Fas in some tissue types. Caspase 8 is a known suppressor of Bcl-2 and activator of “bid,” which also cause cytochrome release. Caspase 8 can also activate caspase 9 if Apaf is absent in the system. (+), Activation; (–), repression; (?), not studied in fetal membrane.
crease in MMP9 levels is seen in the amniotic fluid of women with microbial invasion of the intra-amniotic cavity with and without rupture of membranes; (5) active, inhibitor-free forms of MMP9 are elevated significantly during PROM compared with other conditions; (6) an increased tissue concentration of MMP9 is seen at sites weakened by application of stress to the membranes.12-14, 18-21 Initially, we were confused by the fact that MMP2 is present in much higher concentrations than MMP9 (2125 ng/mL vs 15 ng/mL). However, we were surprised to see that active MMP9 is actually present in higher quantities than active MMP2 (only 0.01% of total MMP2 was active). Other MMPs, such as MMP3 (stromelysin 1) and MMP14 (MT1-MMP), most likely play a role in the activation of other MMPs in addition to having a less direct effect on the membranes themselves (Fig 1). Lei et al16 have shown that MMP release accompanies the activation of apoptosis in a rat model. We hypothesize that apoptosis plays a role in PROM.17 Membranes that have sustained prema-
ture rupture show not only tissue indication of apoptosis, such as DNA fragmentation by TUNEL (terminal deoxynucleotidyl transferase–mediated biotinylated DNA fragment end-labeling assay) assay17 and ligation-mediated PCR, but also molecular indication of the activation of several pathways leading to apoptosis (Fig 2). Interestingly, although expression of pro-apoptotic proteins such as p53, bax, caspase 8, and IL-18 was elevated in PROM compared with PTL, there also seems to be a protection against apoptosis in PTL membranes as indicated by the increase in the anti-apoptotic factor Bcl-2. Bcl-2 transcripts were increased in PTL compared with PROM. In addition to the p53-mediated apoptosis pathway, it also appears that TNF- and Fas-initiated apoptosis is present in PROM. In contrast, this pathway seems incomplete in PTL. The genes of the TNF-mediated pathway including caspase 8, other effector caspases, and TNF receptor–associated death domain are expressed in prematurely ruptured membranes but not in PTL mem-
Volume 184, Number 7 Am J Obstet Gynecol
branes. We have already documented TNF receptor expression in human fetal membranes.22 In summary, it is clear that separate molecular pathways, one leading to PROM and the other to PTL, are activated in the fetal membranes. The initiators for each of the pathways are not known at this time. The factors that tip the balance such that one patient expresses risk as PTL and another as PROM not only have yet to be identified but at present remain so covert that they do not even lead to speculation. REFERENCES
1. American College of Obstetricians and Gynecologists: ACOG Committee. Opinion No. 147. Committee on Obstetrics Practice. Washington: The College ; 1994. 2. Mercer BM, Lewis R. Preterm labor and preterm premature rupture of the membranes. Diagnosis and management. Infect Dis Clin North Am 1997;11:177-201. 3. McGregor JA, French JI. Preterm birth: the role of infection and inflammation. Medscape Womens Health 1997;2:1. 4. 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. 5. Keelan JA, Coleman M, Mitchell MD. The molecular mechanisms of term and preterm labor: recent progress and clinical implications. Clin Obstet Gynecol 1997;40:460-78. 6. Fortunato SJ, Menon RP, Swan KF, Menon R. Inflammatory cytokine (interleukins 1, 6 and 8 and tumor necrosis factor-alpha) release from cultured human fetal membranes in response to endotoxic lipopolysaccharide mirrors amniotic fluid concentrations. Am J Obstet Gynecol 1996;174:1855-61. 7. Mazor M, Chaim W, Horowitz S, Romero R, Glezerman M. The biomolecular mechanisms of preterm labor in women with intrauterine infection. Isr J Med Sci 1994;30:317-22. 8. Gomez R, Ghezzi F, Romero R, Munoz H, Tolosa JE, Rojas I. Premature labor and intra-amniotic infection. Clinical aspects and role of the cytokines in diagnosis and pathophysiology. Clin Perinatol 1995;22:281-342. 9. Draper D, McGregor J, Hall J, Jones W, Beutz M, Hine 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. 10. Vadillo-Ortega F, Gonzalez-Avila G, Selman M, Karchmer S, Meraz N, Ayala A. Collagen metabolism in premature rupture of membranes. Obstet Gynecol 1990;75:84-8. 11. Vadillo-Ortega F, Hernandez A, Gonzalez-Avila G, Bermejo L, Iwata K, Strauss JF. Increased matrix metalloproteinase activity and reduced tissue inhibitor of metalloproteinase-1levels in amniotic fluids from pregnancies complicated by premature rupture of membranes. Am J Obstet Gynecol 1996;174:1371-6. 12. Maymon E, Romero R, Pacora P, Gervasi MT, Gomez R, Edwin SS, et al. Evidence of in vivo differential bioavailability of the active forms of matrix metalloproteinases 9 and 2 in parturition, spontaneous rupture of membranes, and intra-amniotic infection. Am J Obstet Gynecol 2000;183:887-94. 13. Fortunato SJ, Menon R, Lombardi SJ. Collagenolytic enzymes (gelatinases) and their inhibitors in human amniochorionic membranes. Am J Obstet Gynecol 1997;177:731-41. 14. Fortunato SJ, Menon R, Lombardi SJ. Amniochorion gelatinase/gelatinase inhibitor imbalance in vitro: a possible infectious pathway to rupture. Obstet Gynecol 2000;95:240-4. 15. Fortunato SJ, Menon R, Lombardi SJ. MMP/TIMP imbalance in amniotic fluid during PROM: an indirect support for endogenous pathway to membrane rupture. J Perinat Med 1999;27:362-8. 16. Lei H, Furth EE, Kalluri R, Chiou T, Tilly KI, Tilly JL, et al. A program of cell death and extracellular matrix degradation is activated in the amnion before the onset of labor. J Clin Invest 1996;98:1971-8.
Fortunato and Menon 1405
17. Fortunato SJ, Menon R, Lombardi SJ. Programmed cell death (apoptosis): a possible pathway to metalloproteinase activation and fetal membrane degradation in PROM. Am J Obstet Gynecol 2000;182:168-72. 18. Athayde N, Romero R, Gomez R, Maymon E, Pacora P, Mazor M, et al. MMPs-9 in preterm and term human parturition. J Matern Fetal Med 1999;8:213-9. 19. Athayde N, Edwin SS, Romero R, Gomez R, Maymon E, Pacora P, et al. A role for matrixmetalloproteinase-9 in spontaneous rupture of the fetal membranes. Am J Obstet Gynecol 1998; 179:1248-53. 20. Vadillo-Ortega F, Gonzalez-Avila G, Furth EE, Lei H, Muschel RJ, Stetler-Stevenson WG, Strauss JF 3rd. 92-kd type IV collagenase (MMP-9) activity in human amniochorion increases with labor. Am J Pathol 1995;146:148-56. 21. Uchide K, Ueno H, Inoue M, Sakai A, Fujimoto N, Okada Y. Matrix metalloproteinase-9 and tensile strength of fetal membranes in uncomplicated labor. Obstet Gynecol 2000;95:851-5. 22. Fortunato SJ, Menon R, Swan KF. Expression of TNF-alpha and TNFR p55 in cultured amniochorion. Am J Reprod Immunol 1994;32:188-93.
Discussion DR TARIQ A. SIDDIQI, Cincinnati, Ohio. This is a very interesting paper by Fortunato and Menon. PTL and especially preterm PROM are major contributors to overall neonatal morbidity and mortality in the United States. The mechanism or mechanisms responsible for these adverse events remain unknown, despite the concerted efforts of numerous investigators. The fetal membranes are unique in that under normal circumstances they are able to maintain their integrity and withstand the stretch imposed on them by the growing fetus and increasing intrauterine volume. From a structural perspective, the human amniochorion is composed of 8 layers as described by Bourne in 1960, with 5 of these layers comprising the amnion, and 3, the chorion.1 The first layer is the layer of amnion epithelium, which rests on a defined basement membrane. Beneath the basement membrane sequentially lie the compact, fibroblast, and spongy layers. These make up the amnion. The innermost layer of the chorion is the reticular layer, followed by the pseudobasement membrane and trophoblast layers. The major structural components of these layers are cells and extracellular matrix, the latter being both synthesized and degraded by the cellular components. The extracellular matrix, which is made up of a variety of both collagens and noncollagen components such as microfibrils, elastin, laminin, and fibronectin, is believed to provide tensile strength to the membranes especially through type I and III interstitial collagens in the compact layers and type IV collagen in the basement and pseudobasement membranes.2 These collagens, which provide tensile strength and resilience to the amniochorion membranes, are degraded by a variety of collagenases, gelatinases, and stromelysins collectively belonging to the MMP family of enzymes. A variety of unifying theories have been proposed, suggesting broadly that mechanical distension (at term) and infection (when preterm) activates these metalloproteinases, either through loss of cell-matrix contacts and an autocrine-paracrine mechanism or through inflammatory cytokines, respectively resulting in rupture of mem-
1406 Fortunato and Menon
branes.3 Fortunato and Menon measured the expression of specific metalloproteinases and their activators and inhibitors, as well as their amniotic fluid levels, together with the expression of pro-apoptotic and anti-apoptotic genes in two separate cohorts of patients with PTL and preterm PROM and showed statistically significant differences. My questions for these investigators are as follows: 1. What was the a priori hypothesis for this study? 2. How far from the point of membrane rupture was the amniochorion collected and are the changes in MMPs and other factors uniformly distributed across the entire amniochorion or are they focal? 3. How long were the membranes ruptured? 4. Why do the investigators assume that the changes described are causative of rather than resulting from membrane rupture? 5. Why did the investigators use 2 different populations for tissue and amniotic fluid measurements of MMPs and other factors? 6. How do the investigators explain there being no association of MMP expression with infection in light of their own previous work and that of others? REFERENCES
1. Bourne GL. The microscopic anatomy of the human amnion and chorion. Am J Obstet Gynecol 1960;79:1070-3. 2. Malak TM, Ockleford CD, Bell SC, Dalgleish R, Bright N, MacVicar J. Confocal immunofluorescence localization of collagen Types I, III, IV, V and VI and their ultrastructural organization in term fetal membranes. Placenta 1993;14:385-406. 3. Bryant-Greenwood GD. The extracellular matrix of the human fetal membranes: structure and function. Placenta 1998;19:1-11.
DR FORTUNATO (Closing). In answer to Dr Siddiqi’s question, the null hypothesis for this study would be that there is no difference between preterm labor and premature rupture of the membranes. Obviously, the markers tested in this study have proven that null hypothesis wrong. In answer to the question about the point of membrane rupture and the site of the amniochorion collection, the membranes used in this study were not collected from the rupture site. Membranes were randomly chosen from different areas; however, decidua was cleansed off to obtain pure amniochorion. Study is underway to examine the differences, if any, in the expression pattern of the genes (reported in this article) at various regions of the membranes (placental, mid, and cervical regions). In our organ explant system, we have not seen any differences in the expression pattern of cytokine or MMP
June 2001 Am J Obstet Gynecol
mRNAs regardless of the site of collection. All amniochorion (placental, mid, or cervical) in culture achieves a baseline expression after a 48-hour incubation. How long were the membranes ruptured? The average length of time for our patients with PROM with respect to latency is somewhere between 3 and 4 weeks. It is possible that the data would be different if the membranes were collected at the time that rupture was diagnosed or at the end of the cycle during which the patients were treated. However, we have seen that the amniotic fluid that we collect at the time of rupture really differs from or supports the data that we obtain at the tissue level. We see the same relationship in amniotic fluid as we do in tissue. Amniotic and chorionic membranes and amniotic fluid were collected from the same population. They are collected frequently at different times because we do amniocentesis for PROM to rule out infection early on when patients come in. We collected the membranes at the time of delivery. There is an association of MMP expression with infection. In this study we were not able to show the effect of infection in the amniotic fluid MMP levels. This is simply because we have a small sample size. Maymon et al1 have documented increased MMP levels in the amniotic fluid during infection. Here we are comparing the levels of these MMPs between PTL and PROM. Now, there is an effect of infection. We have been able to show that in our previous studies.2 In this report, the affect of infection is not as large as the effect of the differences between the 2 different groups. (This is most likely due to low number of amniotic fluid samples with positive culture). In a separate report, we have documented the effect of TNF on apoptosis. We hypothesize that infection may play a role in the induction and activation of MMPs. Infection may also play role in inducing apoptosis. This review is conducted to point out some early findings in our laboratory that document the existence of 2 separate molecular pathways during PROM and PTL. Further studies are required to thoroughly analyze factors that are common and are differentially present in PROM and PTL. REFERENCES
1. Maymon E, Romero R, Pacora P, Gervasi MT, Gomez R, Edwin SS, et al. Evidence of in vivo differential bioavailability of the active forms of matrix metalloproteinases 9 and 2 in parturition, spontaneous rupture of membranes, and intra-amniotic infection. Am J Obstet Gynecol 2000;183:887-94. 2. Fortunato SJ, Menon R, Lombardi SJ. Amniochorion gelatinase/gelatinase inhibitor imbalance in vitro: a possible infectious pathway to rupture. Obstet Gynecol 2000;95:240-4.
Key words: Premature rupture of membranes, fetal membranes, apoptosis, preterm labor, cytokines, matrix metalloproteinases
Preterm birth complicates 10% to 15% of all pregnancies, and it accounts for high fraction of the perinatal morbidity and mortality rate.1 Spontaneous preterm From The Perinatal Research Center of The Women’s Health Research and Education Foundation, The Women’s Hospital at Centennial Medical Center, Nashville, Tennessee. Supported in part by grants from Columbia/HCA Healthcare Foundation. Presented at the Sixty-eighth Annual Meeting of The Central Association of Obstetricians and Gynecologists, Chicago, Illinois, October 18-21, 2000. Reprint requests: Stephen J. Fortunato, MD, The Perinatal Research Center of the Women’s Health Research and Education Foundation, 2300 Patterson St, Nashville, TN 37203. Copyright © 2001 by Mosby, Inc. 0002-9378/2001 $35.00 + 0 6/6/115122 doi:10.1067mob.2001.115122
birth, caused by preterm labor (PTL) (contractions before 36 weeks’ gestation) or preterm premature rupture of the membranes (PROM) (membrane rupture before the onset of labor) or both account for ~80% of preterm deliveries.1 PROM is associated with 50% of the PTL cases. The frequency of premature birth has actually increased over the past 3 decades, which in part can be attributed to the fact that the exact cause of these conditions is still unknown. The etiology of PTL and PROM is multifactorial, and most of the causes are still unknown. Microbial invasion of the intra-amniotic cavity and infection of the intraamniotic cavity are associated with a majority of these cases. Infection induces a maternal and fetal inflammatory response (histologic chorioamnionitis), increases 1399
1400 Fortunato and Menon
production and release of the inflammatory cytokines (interleukin [IL]-1, IL-6, IL-8, and tumor necrosis factor-α [TNF-α]), which induce prostaglandin production leading to cervical ripening and prostaglandin-mediated contractility. Elevations of these biologic messengers (cytokines and prostaglandins) are considered markers of PTL and PROM. These factors are common to both PTL and PROM and are elevated in the amniotic fluid even in the absence of microbial invasion of the intra-amniotic cavity and infection of the intra-amniotic cavity when compared with term labor.2-8 The puzzling question we address here is why some women experience PROM and some experience PTL with no rupture of membranes when the etiologic factors associated with both these pathologic complications are the same. Infection, smoking, socioeconomic factors, racial differences, nutritional status, multiple gestation, polyhydramnios, and other conditions have all been associated with both PTL and PROM. To date, studies had evaluated the markers that are commonly elevated in both PTL and PROM. A better understanding of the similarities and differences between the pathways leading to each of these conditions may open new avenues for research and treatment. It has been known for a long time that the inflammatory cytokines are elevated in both PTL and PROM and are most likely instrumental in promoting uterine contractility through increased prostaglandin release. More recently, studies from Draper et al,9 Vadillo-Ortega et al,10, 11 Maymon et al,12 and our laboratory13 have shown that intrinsic activation of matrix metalloproteinases (MMPs) is involved in PROM.14, 15 The presence or absence of apoptosis also appears to represent a significant difference between these processes.15, 16 In this article we review differences in the molecular events associated with PROM and PTL. Methods Amniochorionic membranes were collected from the following groups of women at the time of cesarean delivery for obstetric indications. Group 1 consisted of women with PROM who were delivered by cesarean before term (n = 10). Group 2 consisted of women with PTL and no rupture of membranes (n = 10) who were delivered by cesarean before term. A cross-sectional study was designed to evaluate the changes in the levels of MMPs and cytokines in the amniotic fluid during PROM and PTL. Amniotic fluid samples from eligible women were selected from our bank of previously collected and frozen samples. This study included women with preterm PROM, which was defined as spontaneous amniorrhexis before labor and before 37 weeks’ gestation. A positive Nitrazine test result, a positive fern test result, or positive pooling and oligohydramnios on ultrasound examinations diagnosed membrane rupture. The second group consisted
June 2001 Am J Obstet Gynecol
of women with PTL and no rupture of membranes before 37 weeks’ gestation. Women with other medical or obstetric complications were excluded from this study. All amniotic fluid samples were collected by transabdominal amniocentesis under continuous ultrasonographic guidance or through uterine incisions at the time of cesarean delivery. Samples were centrifuged at 1700 rpm for 10 minutes, and the supernatants were stored at –70ºC until analysis. Amniotic fluid samples were also subjected to culture for microorganisms. Only a very few samples were positive for microorganisms (3 in each group: group B streptococci, 3, Klebsiella oxytoca, 1; Staphylococcus aureus, 1; and Candida species, 1). Accordingly, in this report culturepositive and culture-negative sample groups were combined. Histologic examination of the fetal membranes was performed to document histologic chorioamnionitis (>3-4 neutrophils/high-power field), and the positive membranes were excluded from this study. Ribonucleic acid (RNA) extraction and complementary deoxyribonuecleic acid (DNA) synthesis. Total RNA was extracted by the Trizol (Life Technologies, Bethesda, Md) method, and 0.5 µg of total RNA was subjected to oligo dT primed reverse-transcriptase reaction. Complementary deoxyriblonucleic acid (DNA) was then subjected to quantitative polymerase chain reaction (QPCR) or multiple PCR with target-specific primers. QPCR for MMP2, MMP9, MT1-MMP, TIMP2, p53, bax, and Bcl-2. In the competitive PCR assay, one set of primers is used to amplify both the target gene complementary DNA and another neutral DNA fragment. A known concentration of neutral DNA fragment competes with the target cDNA fragment (unknown) for the same primers (primer sequences used in this study were published earlier) and acts as an internal standard. Detailed protocols can be found elsewhere.14, 16 Gene expression of MMPs, TIMP2, pro-apoptotic genes (p53 and bax) and anti-apoptotic (Bcl-2) factors were analyzed in PROM and PTL membranes by QPCR. A constant amount of cDNA was subjected to QPCR. Quantitation of QPCR gels was done by using densitometric analysis with the Alpha-Ease (AlphaInnotech Corp, San Leandro, Calif) software program.14,17 Enzyme-linked immunosorbent assay (ELISA) for MMPs 2, 3, and 9, TIMP2, and interleukin-18. The levels of MMPs (immunoreactive MMP2, 9 [gelatinases], MMP3 [stromelysin 1], and TIMP2) and the pro-apoptotic cytokine, IL-18, were quantitated in the amniotic fluid by means of ELISA. The assay procedure involved a multiple-site 2-step sandwich immunoassay with oligoclonal antibodies. These assays were performed in our laboratory with commercially available kits (Amersham-Pharmacia, Piscataway, NJ [for MMP 2, 3, 9 and TIMP2] R&D Systems, Minneapolis, Minn [for IL-18]). The manufacturer’s instructions were followed for each kit to perform
Fortunato and Menon 1401
Volume 184, Number 7 Am J Obstet Gynecol
Table I. Quantitative expression of MMPs in PROM and PTL by QPCR PROM (n = 8)
MMP2 MMP9 MT1-MMP TIMP2
PTL (n = 8)
Median
±SD
Range
Median
±SD
Range
P value
60,000 33 600 550,000
29,88 0.3 196 287,838
6000-600,000 0.6-60 65-662 55,000-660,000
600 0.6 0.6 535,000
2548 2.9 36.5 252,523
600-6000 0.6-6 0.6-80 50,000-640,000
.001 .02 .006 .9
Data are presented as number of transcripts per 0.5 µg of total RNA.
ELISA. Standard curves were developed by using duplicate samples of known quantities of recombinant proteins. Sample concentrations were determined by relating the absorbance obtained to the standard curve by linear regression analysis. Colorimetric absorption was read at 450 nm by using a microplate reader. Controls consisted of assay buffer. The sensitivities of various assays were as follows: MMP2, 0.37 ng/mL; MMP3, 1.175 ng/mL; MMP9, 0.2 ng/mL; and TIMP2, 1.5 ng/mL. Active MMP9 in amniotic fluid. The amniotic fluid concentration of inhibitor-free, active MMP9 was tested by using previously described methods.15 A novel protein, engineered to contain an MMP9-specific cleavage site, was used as substrate in this assay (Amersham-Pharmacia). The cleavage products were colorimetrically quantitated in an ELISA format after capture onto the wells by incubating at 4ºC overnight in anti-MMP9 antibody– coated microtiter plates. After washing, the plates were incubated at 37ºC with the substrate protein and a detection enzyme for 6 hours. The resulting color was read at 405 nm by using an ELISA plate reader. Quantitation of zymograms for active MMP2. Amniotic fluid samples were subjected to gelatin zymography along with human recombinant active MMP2 standard (Oncogene Research Products, Cambridge, Mass). Different standard dilutions (range, 1 ng–0.0001 ng) were made and subjected to zymography. The quantification of the bands produced by these standards of known concentrations was used (for standard curve) to calculate the concentrations of active MMP2 (a 66-kd band) in amniotic fluid samples with the use of the Alpha-Ease software program.15 Multiplex PCR assay for Fas-FasL apoptotic pathway genes. Multiplex PCR (MPCR) was used to study the expression pattern of pro-apoptotic genes such as Fas, Fas ligand, caspase 8, and Fas-associated death domain in PROM and PTL amniochorionic membranes. In multiple PCR, several target genes are amplified simultaneously in the same reaction tube (Cytoexpress, Multiplex PCR, BioSource International, Camarillo, Calif). Multiplex primers are used to amplify different target transcripts at the same time. In this assay, Fas, Fas ligand, caspase 8, TNF receptor–associated death domain, and Fas-associated death domain gene expression was studied. Glyceralde-
hyde 3-phosphate dehydrogenase was used as the control gene. Complementary DNA from reverse transcriptase reactions was subjected to MPCR. Manufacturer’s instructions were followed for MPCR. DNA fragmentation analysis by ligand-mediated PCR. The extent of DNA damage during PROM and PTL caused by apoptosis of the cells was investigated by using the ligation-mediated PCR assay. The details of this assay protocol are described elsewhere.17 Statistical analysis. Unpaired nonparametric, 2-sample Mann-Whitney U test was used to analyze statistical significance of QPCR and ELISA (P < .05 was considered significant). Results The mean gestational age of fetuses in the PROM group was 31.6 ± 3.4 weeks and of the PTL group was 30 ± 4.3 weeks (P = 0.6). Histologic chorioamnionitis was documented in 40% (4/10) of prematurely ruptured membranes and 50% (5/10) of PTL membranes. The presence of histologic chorioamnionitis was not associated with a statistically significant difference in any of the factors studied in this report. Increased expression of MMPs in PROM. Amniochorionic membranes showed an increased expression of gelatinases (MMP2 and MMP9) and the MMP2 activator protein gene MT1-MMP (MMP14) in membranes from the PROM group compared with those from the PTL group (Table I). In prematurely ruptured membranes, 2.4 × 105 transcripts of MMP2 were seen, whereas only 1.8 × 103 transcripts were seen in PTL membranes. MT1MMP transcripts were similarly increased in prematurely ruptured membranes compared with PTL membranes (545.6 vs 20.9 transcripts, respectively; P = 0.006). MMP9 transcripts were increased from 2.4 in PTL membranes to 31.6 in prematurely ruptured membranes (P = .02). The MMP inhibitor TIMP2 messenger RNA levels were similar in PROM and PTL. The number of transcripts of TIMP2 was 3.5 × 105 in PROM compared with 4.2 × 105 in PTL (P = not significant). MMP and TIMP2 concentrations in amniotic fluid. Immunoreactive MMP9 concentration was significantly increased in the amniotic fluid samples obtained from women with PROM (PTL, 3.3 ng/mL; PROM, 15.25
1402 Fortunato and Menon
June 2001 Am J Obstet Gynecol
Table II. Amniotic fluid levels of MMPs and TIMP2 (n = 18) PROM
MMP2 (ng/mL) MMP9 (ng/mL) MMP 3 (ng/mL) A-MMP2* (pg/mL) A-MMP9* (pg/mL) TIMP2 (ng/mL) IL-18* (pg/mL)
PTL
Median
±SD
Range
Median
±SD
Range
P value (PROM vs PTL)
2400 8.6 1.5 200.5 200 85 1990
420 14.3 29.4 161.7 1130 43.7 3465.7
1300-2500 0-30.9 0-102.9 122-851 0-4900 60-250 710-1522
2050 0 1.1 105.5 0 184 888
879.8 9.5 17.16 85.0 200 130.4 693.4
170-2400 0-9.7 0-69.8 0-263 0-1000 100-560 170-2746
.1 .002 .8 .01 .04 .0001 .001
A, Active levels.
Table III. Quantitative expression of apoptotic factors during PROM and PTL PROM (n = 10)
PTL (n = 10)
Median
±SD
Range
Median
±SD
Range
P value
60 486.5 2.8
25.1 230.4 2.5
6-86 56-892 0-6
3 32 58
17.9 200.3 287.69
0.6-60 0.5-509 0-656
.001 .003 .05
p53 Bax Bcl-2
Data are presented as number of transcripts per 0.5 µg of total RNA.
Table IV. Expression of Fas-FasL–mediated apoptotic pathway genes during PROM and PTL
PROM PTL
Fas
Fas L
Caspase 8
FADD
TRADD
GAPDH
+ ±
– –
+ –
+ +
+ ±
+ +
FADD, Fas-associated death domain; TRADD, TNF receptor– associated death domain; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; ±, expression seen in at least 50% of the tissues.
ng/mL; P = 0.002). In fact, immunoreactive MMP9 was present in detectable concentrations in only 27% (5/18) of the PTL samples, whereas 72% (13/18) of the PROM samples contained detectable immunoreactive MMP9. Although both MMP2 and MMP3 amniotic fluid concentrations were increased in PROM, this was not a statistically significant finding, in part because of the large standard deviation associated with each group (Table II). A significant and inverse relationship was seen in TIMP2 concentrations when compared with the MMPs. The concentration of this MMP inhibitor was decreased in PROM (97.2 ng/mL) when compared with PTL (248.9 ng/mL; P = 0.0001). The molar ratios between MMP2 and its inhibitor/activator TIMP2 were also calculated. The ratio of MMP2 to TIMP2 was 2 in the amniotic fluid of the PTL group compared with 6.3 in the PROM group. TIMP2 was chosen for investigation because it performs the role of MMP2 activator at low equimolar concentrations and the role of MMP2 inhibitor at higher concentrations. The active, TIMP-free forms of MMP2 and 9 in amniotic fluid were assayed. Less than 1% of the total proteins were active in both PROM and PTL, and the active forms
were elevated in PROM compared with PTL. Zymogram quantitation of MMP2 revealed an active MMP2 level of 233.5 pg/mL in the amniotic fluid samples from PROM, whereas the mean level was 127.8 pg/mL in PTL (P = .01). Amniotic fluid active MMP9 levels were 600 pg/mL in PROM samples, whereas the levels were below the detection limits of the assay in most (83%) of the PTL samples (P = .04). Of note, Athayde et al18, 19 have reported active MMP9 in the amniotic fluid of women with PTL with intact membranes using an assay modified to detect lower concentrations. Their data are in agreement with our findings that active MMP9 levels are higher in PROM compared with PTL.12 A specific activity assay for MMP3 was not performed because such a method is not yet commercially available. Elevation of apoptotic elements during PROM. It is hypothesized that apoptosis may be responsible for the MMP increase in PROM because some pro-apoptotic proteins such as p53 are capable of transactivating some of the MMP genes.17 DNA fragmentation (intra-nucleosomal cleavage and laddering of DNA) (data not shown) was noticed in the majority of the prematurely ruptured membranes (8/10) compared with 30% (3/10) of PTL membranes.17 We quantitated the proapoptotic protein p53 induced by DNA damage, as well as p53 target proteins such as bax (pro-apoptotic) and Bcl-2 (anti-apoptotic), using QPCR (Table III). Compared with PTL, the expression of pro-apoptotic genes was elevated in the prematurely ruptured membranes. p53 transcripts were increased from 9.3 in PTL membranes to 49.6 (P = .003) in prematurely ruptured membranes, and bax transcripts were 120.8 and 524.3 in PTL and PROM, respectively (P = .004). The number of anti-
Volume 184, Number 7 Am J Obstet Gynecol
Fortunato and Menon 1403
Fig 1. Factors associated with PROM can increase expression of MT1-MMP, MMP2, and MMP3 from human fetal membranes and induce MMP9 expression or release. These factors reduce MMP inhibitor TIMP2. MT1-MMP and low levels of TIMP2 activate pro-MMP2 to active form. Active MMP2 and 3, along with various other proteases, activate MMP9. All these active MMPs can degrade extracellular matrix proteins to cause membrane weakness and rupture. Factors of MMP activation can also activate the apoptotic pathway mediated by p53 and TNF (see Fig 2). p53 is a known inducer of the MMP2 gene. Synergistic action of all these factors may activate a vicious cycle of events resulting in membrane rupture.
apoptotic protein Bcl2 gene transcripts was 100-fold higher in PTL compared with that of PROM (2.8; P = .05), suggesting predominance of an anti-apoptotic effect in PTL. Existence of Fas- and TNF-mediated apoptotic pathway. MPCR showed the expression of TNF- and Fas-mediated apoptotic pathway genes in PROM membranes (Table IV). Fas, an apoptosis inducer, was seen in the majority of prematurely ruptured membranes (9/10) and some PTL membranes (5/10). Caspase 8, an activator/initiator of the caspase cascade of cell death proteins, was expressed in a majority of the prematurely ruptured membranes (9/10), whereas only 30% (3/10) of PTL membranes showed caspase 8 gene expression. The docking proteins of the Fas- and TNF-mediated apoptotic signaling pathway (Fas-associated death domain; TNF-associated death domain) were present in many of the PTL (5 of 10) and prematurely ruptured (7 of 10) membranes. Fas ligand expression was never seen in fetal membranes during either condition. Glyceraldehyde 3-phosphate dehydrogenase was used as a “housekeeping gene” in this experiment and was present in all the membranes tested. The level of the pro-apoptotic and pro-inflammatory cytokine, IL-18, was also elevated in the amniotic fluid during PROM. The mean levels were 1067 pg/mL in PTL compared with 3230 pg/mL in PROM (P = .001). Comment In analyzing the data presented herein, it is obvious that distinct pathways leading to each of the conditions
studied (PROM and PTL) exist. Although it would seem obvious that this should be the case, there is a paucity of data addressing this area. The existence of this separation is made less obvious by the fact that each of these conditions has exactly the same set of etiologic factors. Although PROM and PTL frequently accompany each other, it is well known that they can present as totally separate entities. Both pathways are known to share activation of the cytokine pathway to a greater or lesser degree, it seems, depending on the presence or absence of infection. We have been able to identify several branch points that differentiate one from the other. Initially, our laboratory began looking at the role of the MMPs in PROM. It is now clear that the fetal membranes are endowed with the full system of MMPs, MMP inhibitors (TIMPs), MMP activators, and regulatory elements. Most of the MMPs are increased in membranes that have sustained premature rupture. The active forms of these MMPs are increased in PROM when compared with any other conditions in pregnancy. Several patterns of data lead us to speculate that active MMP9 (gelatinase B) plays a crucial role in PROM: (1) the MMP9 gene is induced in amniochorion during labor, PROM, and infection; (2) MMP9 levels are increased in the amniotic fluid as gestation progresses and also during active labor at term and before term; (3) gelatinolytic activity, corresponding to latent and active forms of MMP9, is increased in amniotic fluid obtained from patients with PROM; (4) a significant in-
1404 Fortunato and Menon
June 2001 Am J Obstet Gynecol
Fig 2. Two major apoptotic pathways may exist in PROM, which can be initiated by infection, genotoxic agents, or unknown factors: (1) TNF receptor–Fas–mediated pathway—these receptor proteins bind to their respective ligands TNF and Fas L, which initiate signal transduction through 2 docking proteins TRADD (TNF receptor– associated death domain) and FADD (Fas-associated death domain). These death domain proteins activate procaspase 8 to active caspase 8. (2) p53-mediated pathway is initiated by DNA fragmentation. DNA damage increases transactivator protein p53 in the cell. p53 transactivates bax, which causes damage to the mitochondrial membrane, resulting in the release of cytochrome c. Cytochrome c activates Apaf (apoptosis protease activating factor), which converts pro-caspase 9 to active form. p53 also suppresses Bcl-2, a factor that stops mitochondrial membrane damage. Active caspase 8 or 9 can initiate a cascade of caspase activation. Caspases 3, 7, and 6 are activated sequentially, which will cause proteolysis of structural proteins, proteins of homeostasis, and several other proteins and program the cell to death. These pathways cross over at several points. p53 can turn on Fas in some tissue types. Caspase 8 is a known suppressor of Bcl-2 and activator of “bid,” which also cause cytochrome release. Caspase 8 can also activate caspase 9 if Apaf is absent in the system. (+), Activation; (–), repression; (?), not studied in fetal membrane.
crease in MMP9 levels is seen in the amniotic fluid of women with microbial invasion of the intra-amniotic cavity with and without rupture of membranes; (5) active, inhibitor-free forms of MMP9 are elevated significantly during PROM compared with other conditions; (6) an increased tissue concentration of MMP9 is seen at sites weakened by application of stress to the membranes.12-14, 18-21 Initially, we were confused by the fact that MMP2 is present in much higher concentrations than MMP9 (2125 ng/mL vs 15 ng/mL). However, we were surprised to see that active MMP9 is actually present in higher quantities than active MMP2 (only 0.01% of total MMP2 was active). Other MMPs, such as MMP3 (stromelysin 1) and MMP14 (MT1-MMP), most likely play a role in the activation of other MMPs in addition to having a less direct effect on the membranes themselves (Fig 1). Lei et al16 have shown that MMP release accompanies the activation of apoptosis in a rat model. We hypothesize that apoptosis plays a role in PROM.17 Membranes that have sustained prema-
ture rupture show not only tissue indication of apoptosis, such as DNA fragmentation by TUNEL (terminal deoxynucleotidyl transferase–mediated biotinylated DNA fragment end-labeling assay) assay17 and ligation-mediated PCR, but also molecular indication of the activation of several pathways leading to apoptosis (Fig 2). Interestingly, although expression of pro-apoptotic proteins such as p53, bax, caspase 8, and IL-18 was elevated in PROM compared with PTL, there also seems to be a protection against apoptosis in PTL membranes as indicated by the increase in the anti-apoptotic factor Bcl-2. Bcl-2 transcripts were increased in PTL compared with PROM. In addition to the p53-mediated apoptosis pathway, it also appears that TNF- and Fas-initiated apoptosis is present in PROM. In contrast, this pathway seems incomplete in PTL. The genes of the TNF-mediated pathway including caspase 8, other effector caspases, and TNF receptor–associated death domain are expressed in prematurely ruptured membranes but not in PTL mem-
Volume 184, Number 7 Am J Obstet Gynecol
branes. We have already documented TNF receptor expression in human fetal membranes.22 In summary, it is clear that separate molecular pathways, one leading to PROM and the other to PTL, are activated in the fetal membranes. The initiators for each of the pathways are not known at this time. The factors that tip the balance such that one patient expresses risk as PTL and another as PROM not only have yet to be identified but at present remain so covert that they do not even lead to speculation. REFERENCES
1. American College of Obstetricians and Gynecologists: ACOG Committee. Opinion No. 147. Committee on Obstetrics Practice. Washington: The College ; 1994. 2. Mercer BM, Lewis R. Preterm labor and preterm premature rupture of the membranes. Diagnosis and management. Infect Dis Clin North Am 1997;11:177-201. 3. McGregor JA, French JI. Preterm birth: the role of infection and inflammation. Medscape Womens Health 1997;2:1. 4. 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. 5. Keelan JA, Coleman M, Mitchell MD. The molecular mechanisms of term and preterm labor: recent progress and clinical implications. Clin Obstet Gynecol 1997;40:460-78. 6. Fortunato SJ, Menon RP, Swan KF, Menon R. Inflammatory cytokine (interleukins 1, 6 and 8 and tumor necrosis factor-alpha) release from cultured human fetal membranes in response to endotoxic lipopolysaccharide mirrors amniotic fluid concentrations. Am J Obstet Gynecol 1996;174:1855-61. 7. Mazor M, Chaim W, Horowitz S, Romero R, Glezerman M. The biomolecular mechanisms of preterm labor in women with intrauterine infection. Isr J Med Sci 1994;30:317-22. 8. Gomez R, Ghezzi F, Romero R, Munoz H, Tolosa JE, Rojas I. Premature labor and intra-amniotic infection. Clinical aspects and role of the cytokines in diagnosis and pathophysiology. Clin Perinatol 1995;22:281-342. 9. Draper D, McGregor J, Hall J, Jones W, Beutz M, Hine 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. 10. Vadillo-Ortega F, Gonzalez-Avila G, Selman M, Karchmer S, Meraz N, Ayala A. Collagen metabolism in premature rupture of membranes. Obstet Gynecol 1990;75:84-8. 11. Vadillo-Ortega F, Hernandez A, Gonzalez-Avila G, Bermejo L, Iwata K, Strauss JF. Increased matrix metalloproteinase activity and reduced tissue inhibitor of metalloproteinase-1levels in amniotic fluids from pregnancies complicated by premature rupture of membranes. Am J Obstet Gynecol 1996;174:1371-6. 12. Maymon E, Romero R, Pacora P, Gervasi MT, Gomez R, Edwin SS, et al. Evidence of in vivo differential bioavailability of the active forms of matrix metalloproteinases 9 and 2 in parturition, spontaneous rupture of membranes, and intra-amniotic infection. Am J Obstet Gynecol 2000;183:887-94. 13. Fortunato SJ, Menon R, Lombardi SJ. Collagenolytic enzymes (gelatinases) and their inhibitors in human amniochorionic membranes. Am J Obstet Gynecol 1997;177:731-41. 14. Fortunato SJ, Menon R, Lombardi SJ. Amniochorion gelatinase/gelatinase inhibitor imbalance in vitro: a possible infectious pathway to rupture. Obstet Gynecol 2000;95:240-4. 15. Fortunato SJ, Menon R, Lombardi SJ. MMP/TIMP imbalance in amniotic fluid during PROM: an indirect support for endogenous pathway to membrane rupture. J Perinat Med 1999;27:362-8. 16. Lei H, Furth EE, Kalluri R, Chiou T, Tilly KI, Tilly JL, et al. A program of cell death and extracellular matrix degradation is activated in the amnion before the onset of labor. J Clin Invest 1996;98:1971-8.
Fortunato and Menon 1405
17. Fortunato SJ, Menon R, Lombardi SJ. Programmed cell death (apoptosis): a possible pathway to metalloproteinase activation and fetal membrane degradation in PROM. Am J Obstet Gynecol 2000;182:168-72. 18. Athayde N, Romero R, Gomez R, Maymon E, Pacora P, Mazor M, et al. MMPs-9 in preterm and term human parturition. J Matern Fetal Med 1999;8:213-9. 19. Athayde N, Edwin SS, Romero R, Gomez R, Maymon E, Pacora P, et al. A role for matrixmetalloproteinase-9 in spontaneous rupture of the fetal membranes. Am J Obstet Gynecol 1998; 179:1248-53. 20. Vadillo-Ortega F, Gonzalez-Avila G, Furth EE, Lei H, Muschel RJ, Stetler-Stevenson WG, Strauss JF 3rd. 92-kd type IV collagenase (MMP-9) activity in human amniochorion increases with labor. Am J Pathol 1995;146:148-56. 21. Uchide K, Ueno H, Inoue M, Sakai A, Fujimoto N, Okada Y. Matrix metalloproteinase-9 and tensile strength of fetal membranes in uncomplicated labor. Obstet Gynecol 2000;95:851-5. 22. Fortunato SJ, Menon R, Swan KF. Expression of TNF-alpha and TNFR p55 in cultured amniochorion. Am J Reprod Immunol 1994;32:188-93.
Discussion DR TARIQ A. SIDDIQI, Cincinnati, Ohio. This is a very interesting paper by Fortunato and Menon. PTL and especially preterm PROM are major contributors to overall neonatal morbidity and mortality in the United States. The mechanism or mechanisms responsible for these adverse events remain unknown, despite the concerted efforts of numerous investigators. The fetal membranes are unique in that under normal circumstances they are able to maintain their integrity and withstand the stretch imposed on them by the growing fetus and increasing intrauterine volume. From a structural perspective, the human amniochorion is composed of 8 layers as described by Bourne in 1960, with 5 of these layers comprising the amnion, and 3, the chorion.1 The first layer is the layer of amnion epithelium, which rests on a defined basement membrane. Beneath the basement membrane sequentially lie the compact, fibroblast, and spongy layers. These make up the amnion. The innermost layer of the chorion is the reticular layer, followed by the pseudobasement membrane and trophoblast layers. The major structural components of these layers are cells and extracellular matrix, the latter being both synthesized and degraded by the cellular components. The extracellular matrix, which is made up of a variety of both collagens and noncollagen components such as microfibrils, elastin, laminin, and fibronectin, is believed to provide tensile strength to the membranes especially through type I and III interstitial collagens in the compact layers and type IV collagen in the basement and pseudobasement membranes.2 These collagens, which provide tensile strength and resilience to the amniochorion membranes, are degraded by a variety of collagenases, gelatinases, and stromelysins collectively belonging to the MMP family of enzymes. A variety of unifying theories have been proposed, suggesting broadly that mechanical distension (at term) and infection (when preterm) activates these metalloproteinases, either through loss of cell-matrix contacts and an autocrine-paracrine mechanism or through inflammatory cytokines, respectively resulting in rupture of mem-
1406 Fortunato and Menon
branes.3 Fortunato and Menon measured the expression of specific metalloproteinases and their activators and inhibitors, as well as their amniotic fluid levels, together with the expression of pro-apoptotic and anti-apoptotic genes in two separate cohorts of patients with PTL and preterm PROM and showed statistically significant differences. My questions for these investigators are as follows: 1. What was the a priori hypothesis for this study? 2. How far from the point of membrane rupture was the amniochorion collected and are the changes in MMPs and other factors uniformly distributed across the entire amniochorion or are they focal? 3. How long were the membranes ruptured? 4. Why do the investigators assume that the changes described are causative of rather than resulting from membrane rupture? 5. Why did the investigators use 2 different populations for tissue and amniotic fluid measurements of MMPs and other factors? 6. How do the investigators explain there being no association of MMP expression with infection in light of their own previous work and that of others? REFERENCES
1. Bourne GL. The microscopic anatomy of the human amnion and chorion. Am J Obstet Gynecol 1960;79:1070-3. 2. Malak TM, Ockleford CD, Bell SC, Dalgleish R, Bright N, MacVicar J. Confocal immunofluorescence localization of collagen Types I, III, IV, V and VI and their ultrastructural organization in term fetal membranes. Placenta 1993;14:385-406. 3. Bryant-Greenwood GD. The extracellular matrix of the human fetal membranes: structure and function. Placenta 1998;19:1-11.
DR FORTUNATO (Closing). In answer to Dr Siddiqi’s question, the null hypothesis for this study would be that there is no difference between preterm labor and premature rupture of the membranes. Obviously, the markers tested in this study have proven that null hypothesis wrong. In answer to the question about the point of membrane rupture and the site of the amniochorion collection, the membranes used in this study were not collected from the rupture site. Membranes were randomly chosen from different areas; however, decidua was cleansed off to obtain pure amniochorion. Study is underway to examine the differences, if any, in the expression pattern of the genes (reported in this article) at various regions of the membranes (placental, mid, and cervical regions). In our organ explant system, we have not seen any differences in the expression pattern of cytokine or MMP
June 2001 Am J Obstet Gynecol
mRNAs regardless of the site of collection. All amniochorion (placental, mid, or cervical) in culture achieves a baseline expression after a 48-hour incubation. How long were the membranes ruptured? The average length of time for our patients with PROM with respect to latency is somewhere between 3 and 4 weeks. It is possible that the data would be different if the membranes were collected at the time that rupture was diagnosed or at the end of the cycle during which the patients were treated. However, we have seen that the amniotic fluid that we collect at the time of rupture really differs from or supports the data that we obtain at the tissue level. We see the same relationship in amniotic fluid as we do in tissue. Amniotic and chorionic membranes and amniotic fluid were collected from the same population. They are collected frequently at different times because we do amniocentesis for PROM to rule out infection early on when patients come in. We collected the membranes at the time of delivery. There is an association of MMP expression with infection. In this study we were not able to show the effect of infection in the amniotic fluid MMP levels. This is simply because we have a small sample size. Maymon et al1 have documented increased MMP levels in the amniotic fluid during infection. Here we are comparing the levels of these MMPs between PTL and PROM. Now, there is an effect of infection. We have been able to show that in our previous studies.2 In this report, the affect of infection is not as large as the effect of the differences between the 2 different groups. (This is most likely due to low number of amniotic fluid samples with positive culture). In a separate report, we have documented the effect of TNF on apoptosis. We hypothesize that infection may play a role in the induction and activation of MMPs. Infection may also play role in inducing apoptosis. This review is conducted to point out some early findings in our laboratory that document the existence of 2 separate molecular pathways during PROM and PTL. Further studies are required to thoroughly analyze factors that are common and are differentially present in PROM and PTL. REFERENCES
1. Maymon E, Romero R, Pacora P, Gervasi MT, Gomez R, Edwin SS, et al. Evidence of in vivo differential bioavailability of the active forms of matrix metalloproteinases 9 and 2 in parturition, spontaneous rupture of membranes, and intra-amniotic infection. Am J Obstet Gynecol 2000;183:887-94. 2. Fortunato SJ, Menon R, Lombardi SJ. Amniochorion gelatinase/gelatinase inhibitor imbalance in vitro: a possible infectious pathway to rupture. Obstet Gynecol 2000;95:240-4.