Role of matrix metalloproteinases in preterm labour


Dr F. Vadillo-Ortega, Instituto Nacional de Perinatologia, Montes Urales 800, Lomas de Virreyes, Mexico City, 11000, Mexico.


Extracellular matrix homeostasis is a key process in the maintenance of the tensile strength of the amniochorion. This tensile strength guarantees the role of the membranes as a physical and functional boundary for the fetus during human pregnancy. Pathological rupture of these structures before 37 completed weeks of gestation is known as preterm prelabour rupture of the membranes (PPROM) and it is a major cause of spontaneous preterm labour and preterm birth. A mechanism involving the activation of matrix metalloproteinases (MMP)-9, a 92-kDa type IV collagenase, as an essential mediator of tissue damage is under investigation. The proposed mechanism involves the abnormal expression and activity of MMP-9 with subsequent connective tissue degradation taking place at a time that does not synchronise with other events of labour. The local physiological signal by amniochorion cells to induce MMP-9 expression is not known, but bacterial products and/or the proinflammatory cytokines, IL-1β and TNF-α, as paracrine or autocrine signals may trigger these processes in pregnancies complicated with intra-amniotic infection. These signalling pathways indicate complex cooperative and bidirectional communications between amnion and choriodecidua in response to bacterial products, which include intermembranous cytokine traffic and signalling between tissues. Products secreted in culture by amniochorion and choriodecidual leucocytes, obtained from women who delivered following normal labour in the absence of infection, condition a specific microenvironment that induces collagen degradation in fetal membranes. Further characterisation of the role of choriodecidual leucocytes in the control of extracellular matrix degradation in amniochorion is currently under way.


Prelabour rupture of the fetal membranes (PROM) is defined as the transvaginal loss of amniotic fluid in the absence of active labour at any gestational age. This pregnancy complication has clinical relevance when it occurs before 37 weeks of gestation [preterm prelabour rupture of the membrane (PPROM)] and may be related to at least 30% of all cases of spontaneous preterm labour, making PPROM the most commonly identified direct cause of preterm birth. As a single pathology, PPROM may complicate 15% of all pregnancies with a higher prevalence observed in Hispanic and Afro-American women,1 although the underlying mechanisms explaining the development of PROM are still poorly understood and its prevention and treatment remains rudimentary.

Recently, a general hypothesis of the pathophysiological mechanisms of PROM has been proposed.2 Abnormal degradation of the extracellular matrix in the amniochorion has been proposed as the final pathway, leading to membrane rupture in the absence of the mechanical forces generated by the contractile activity of the myometrium.

The histology of the fetal membranes is characterised by the abundant proportion of connective tissue in both the amnion and the chorion. Most of the mechanical tensile strength of these tissues reside in the compact layer of the amnion, its name derived from the tight arrangement of abundant collagen fibres and the absence of cells. The reticular and spongy layers of the chorion also contribute towards mechanical support. Amniochorionic extracellular matrix is composed of several different genetic types of collagen arranged in a complex framework, maximising its mechanical resistance. Major components are types I, III, IV, V and VI collagens and abundant proteoglycans, which are embedded in the fibrous proteins. It is widely accepted that the principal tissue support is generated by fibres composed of types I and III collagens, which themselves are stabilised by a network of collagen types: IV, V and VI.3 These structures lose their architecture and mechanical strength around labour synchronously with increased myometrium activity and cervical ripening, allowing delivery to occur. Prelabour rupture of the fetal membranes is characterised by the abnormal activation of this process and constitutes the clinical evidence that specific independent biochemical mechanisms associated with rupture of the amniochorion can be initiated in the absence of other labour events.


The molecular components of connective tissue are highly resistant to the action of extracellular proteases and physiological degradation requires the participation of a group of specific proteases called matrix metalloproteinases (MMPs). This is a subfamily of metalloproteinases formally known as matrixins, which are composed of at least 23 different proteins with sequences similar to that of the human interstitial collagenase also known as MMP-1. MMPs are secreted in an inactive form and are inhibited by MMP-specific tissue inhibitors.4 Even though the specific role of all MMPs in the degradation of the amniochorionic extracellular matrix has not been described, it is now evident that a number of substances composed of at least gelatinases, collagenases and stromelysin is expressed during amniochorionic rupture. Major enzymes associated with this process are gelatinases A and B (MMP-2 and MMP-9, respectively) and collagenases 2 and 3, also known as MMP-8 and MMP-13. Other MMPs such as stromelysin-1 (MMP-3) are expressed during labour. The sequence of events leading to the degradation of the extracellular matrix has not been fully defined but, according to the substrate affinity of the enzymes mentioned above, it has been proposed that their activity may result in membrane rupture (Table 1).

Table 1.  Substrate affinity of the different MMPs expressed in fetal membranes.
Type I collagenYesNoYesYesYes
Type III collagenYesNoYesYesNo
Type IV collagenYesYesNoNoNo
Type V collagenYesYesNoNoNo
Type VI collagenNoNoNoYesNo

MMP-2 is constitutively expressed on the amnion during gestation in several animal species, including humans. The activity of MMP-2 is enhanced during labour. MMP-9 is selectively expressed at the end of gestation by the amnion, trophoblasts and decidual cells and is the major MMP expressed during labour. A correlation between MMP-9 enzyme expression and the decline of membrane tensile strength has also been described.5 The selective timing of MMP-9 expression makes it a potential candidate as a molecular marker in the initiation of labour. MMP-13 is secreted by the amnion during labour whereas infiltrating polymorphonuclear cells secrete MMP-8. Specific subpopulations of T lymphocytes (CD3+) expressing MMP-9 are attracted to the choriodecidual interphase and they are capable of establishing ‘cross-talking’ with other cell populations located nearby in the membranes and together they collaborate in the degradation of the extracellular matrix.

The general hypothetical model for normal membrane rupture during labour involves the induction of MMP expression, secretion and activation by local and infiltrating cells synchronously with myometrial and cervical changes. This leads to the progressive loss of the mechanical properties of the membranes and eventually to their rupture. Alternatively, PROM can be explained by the abnormal activation of this enzymatic system in such a way that it is not synchronous with other mechanisms of labour. The search for signals that control the expression, secretion and activation of these enzymes in the microenvironment of the amniochorion is consequently an area of great interest.


Several cytokines have been identified as having the potential to modulate MMP expression in fetal membranes (Table 2), but the primary signal leading to activation of these tissues during labour or PROM remains elusive. Current knowledge on the mechanisms of amniochorion connective tissue degradation has led to the proposal that there is participation of several different cell populations, in such a way that the histological layers containing the mechanical support of the extracellular matrix are flanked as a ‘sandwich’ by actively MMP-secreting cells from the amnion and chorion during the initiation of labour. In humans and primates, the secretion of several MMPs occurs five to six days before birth, yet several hours before birth in the rat. This makes the accumulation of the inactive MMP in the extracellular matrix necessary as an accompanying process during the final days of gestation, as well as the existence of an activation mechanism to start the transition between the pre-existing pro-MMP and their active forms during delivery. This mechanism of MMP activation is completely unknown.

Table 2.  Inflammation-related cytokines that modulate MMP expression in amniochorion.
CytokineTargetResulting action
IL-1βAmnion epithelial cellsIncrease in MMP-9 expression and activity
TrophoblastsMMP-9 secretion
TNF-αChorion cellsMMP-9 expression
TrophoblastsMMP-9 increased activity and MMP-2 decreased activity
Amnion cellsIncreased secretion and activity of MMP-9
IL-6TrophoblastsIncreased MMP-9 and MMP-2 activity

In vitro stimulation of the human amniochorion with either interleukin-1β (IL-1β) or tumour necrosis factor-α (TNF-α) results in a dose-dependent secretion of the MMP-9 proenzyme but does not result in activation.6 Functional interfering experiments using entire amniochorion stimulated with lipopolysaccharide (LPS) have shown that IL-1β is the key cytokine for induction of MMP-9 expression in these tissues.7 IL-1β is secreted by the chorion following recognition of pathogen-associated patterns (PAMPs) by local cells, such as macrophages, decidual cells, amnion cells and trophoblasts, all of which express toll-like receptors (TLR).8 IL-1β and several other cytokines appear to have an intramembranous transport route because they can be secreted inside the amnion, appearing in the amniotic fluid, even though the mRNA for IL-1β has only been documented in chorion cells. This observation could be explained by the existence of a currently undefined intercellular pathway, which would guarantee specific cytokine and MMP secretion.9


Enough clinical and experimental data exist to support the causal nature of cervicovaginal and/or intrauterine infections in explaining the development of PPROM and spontaneous preterm labour. The inflammatory response initiated by ascending infection may cause ‘cross-talking’ with reproductive tract tissues, resulting in the activation of mechanisms leading to labour. Despite several signals having been identified, it remains unclear as to how the same infectious process can result in so many alternative outcomes even in genetically similar individuals. In some women, infection can be controlled and pregnancy can continue to term without complications. In others, the main event is activation of myometrial activity, leading to threatening spontaneous preterm labour while in others, the primary result is the secretion of MMP by fetal membranes, leading to PPROM.

Inflammation-related signals, which modulate MMP expression in amniochorionic cells, are summarised in Table 2. From this list, it can be concluded that there is a direct link between infection and PPROM. The use of experimental animal models intended to explore the sequence of events leading to PPROM has shown that the process is more complex than a simple relationship between cytokines and MMP secretion. When analysing choriodecidual infection with group B streptococcus (GBS) in a rhesus monkey model, it was observed that by increasing the amount of live bacteria injected into the choriodecidua, a dose-dependent increase of TNF-α and IL-1β in the amniotic fluid was provoked. This was followed by augmentation in MMP-9 and MMP-2 levels in the same compartment, suggesting degradation initiation of the extracellular matrix in the amniochorion.10 Direct injection of IL-1β into the amniotic fluid was followed, within hours, by a dose-dependent secretion of pro-MMP-9. Although 50% of animals receiving choriodecidual bacteria ended their pregnancies with spontaneous preterm labour, none of the animals were found to develop PPROM. These animals were also positive for GBS in the amniotic fluid. Therefore, further exploration as to the possibility that PPROM development may be due to chronic choriodecidual infections is required.

An individual's capacity to respond to an infectious agent is also a potential alternative in understanding why some individuals progress to PPROM or spontaneous preterm labour whereas others do not. In order to evaluate the role of the genetic background in the development of PPROM, we have explored the in vitro response of fetal membranes obtained from women carrying different polymorphisms of the pro-inflammatory cytokines: TNF-α and IL-1β. One such polymorphism is characterised by a single G>A point mutation at region −308 in the promoter region of TNF-α (also known as polymorphism-2). This polymorphism alteration is known to enhance TNF-α secretion and is correlated with poor clinical outcomes in patients exposed to systemic infections. This also holds true for a +3953 polymorphism-2 of the IL-1β gene, which affects a non-coding region of the gene but also results in increased cytokine secretion. When membranes carrying any of these polymorphisms were exposed in vitro to bacterial LPS, they were shown to secrete more of the corresponding cytokine when compared with membranes carrying the normal, most common polymorphism. A twofold increased secretion of TNF-α was observed only after exposing the membranes to high doses of LPS, although secretion of IL-1β was at least three times higher by the membranes carrying the IL-1*2 allele than the most common polymorphism (IL1*1) at any dose of LPS.11 These augmented cytokine secretions are also correlated with increased MMP-9 and MMP-2 secretion by the same membranes.

This alternative response to infection-related compounds, by tissues carrying different pro-inflammatory cytokine polymorphisms, may be used to highlight the hypothesis that differential clinical outcomes may be expected when pregnancies are exposed to intrauterine infection and the mother carries a polymorphism (i.e. polymorphism-1 of IL-1β) and the fetal tissues carry an allele derived from the father, expressing a hyper-responsive gene (polymorphism-2 of IL-1β). From these biological responses, it could be suggested that when the mother/fetus are exposed to infection, the amniochorion may develop a higher response than the maternal tissues (i.e. myometrium), leading to a preferential activation of the mechanisms that induce the degradation of the molecular components of the extracellular matrix in the amniochorion. As a consequence, PPROM may represent the clinical manifestation of this genetic difference. Alternatively, if the mother is carrying the hyperfunctional allele, the induction of maternal tissue (the myometrium) activity may be the primary response with spontaneous preterm labour appearing preferentially.


More work in the field of MMPs and PPROM must be targeted towards further understanding of the role of these enzymes in the process of abnormal membrane rupture. The identification of MMP-9 and other MMPs as potential biological or pathological mediators in the degradation of connective tissue raises the possibility of the direct inhibition of these enzymes, resulting in a therapeutic target in the management of PPROM. Several specific and non-specific MMP inhibitors are now available and are undergoing clinical testing in the area of cancer metastasis. It is clear that in the near future, clinical management of PPROM may involve the application of these compounds. In addition to the biological role of MMP in PPROM, MMPs have been shown to be suitable clinical markers of amniochorionic activation. As a result, they can be used to predict pregnant woman who may be susceptible to PPROM and/or spontaneous preterm labour.