| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Minireview |
a Department of Obstetrics and Gynecology, University of Kentucky, Lexington, Kentucky 40536-0293
b Women's Reproductive Health Research Center, Vanderbilt University, Nashville, Tennessee 37232-2519
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONOVARYUTERUSSUMMARYREFERENCES |
|---|
menstrual cycle, ovary, ovulation, steroid hormones, uterus
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONOVARYUTERUSSUMMARYREFERENCES |
|---|
In the female reproductive tract, the MMP system has been postulated to regulate many of the cyclic changes that occur in the most dynamic structural components of the ovary and uterus. These changes in tissue architecture, which are orchestrated by various hormones, growth factors, and cytokines, are crucial to follicular/luteal function in the ovary and to endometrial function in the uterus. The present review focuses on the role of the MMP system in the reproductive function of female mammals, highlighting key elements of tissue remodeling in the ovary and uterus.
Metalloproteinases
The MMPs are a family of more than 20 related proteolytic enzymes [1–4] that includes four broad classes: 1) collagenases, 2) gelatinases, 3) stromelysins, and 4) membrane-type enzymes (MT-MMPs), as noted in Table 1 and depicted in Figure 1. These proteinases exhibit numerous similarities in their structure and function (Fig. 1). Common features include: 1) synthesis of the MMPs as preproenzymes that are secreted as inactive pro-MMPs, 2) activation of the latent zymogen in the extracellular space, 3) recognition and cleavage of the ECM by the catalytic domain, and 4) inhibition of MMP action in the extracellular environment by both serum-borne and tissue-derived metalloproteinase inhibitors. Because degradation of the ECM is an exquisitely controlled process, MMP activation is a critical point of regulation [1–3]. Activation of the pro-MMPs in the extracellular space is proposed to occur via proteinases, including other MMPs and the plasminogen-activator/plasmin system, as well as by nonproteolytic agents, such as reactive oxygen species. One notable exception to the extracellular activation of pro-MMPs is MMP-11 also called stromelysin-3. This MMP contains a small region following the propeptide domain that is catalytically cleaved intracellularly by furins, which are protein-processing enzymes. This intracellular cleavage of the propeptide domain results in MMP-11 being secreted in an active form [4].
|
|
Despite many similarities in the structure of the MMPs, distinct differences occur in the recognition and specificity of the various MMPs for components of the ECM [1–3] (Table 1). For example, the collagenases (i.e., MMP-1, MMP-8, and MMP-13) are able to cleave both fibrillar and nonfibrillar collagens. With the fibrillar collagens, collagenases catalyze a crucial cleavage in the triple helical collagen molecule, which changes the collagen stability and solubility properties. The products of collagenase action rapidly denature into gelatin and, thus, become susceptible to a wide variety of tissue proteinases, including the gelatinases and stromelysins. The gelatinases (i.e., MMP-2 and MMP-9) contain a fibronectin-like sequence within their catalytic domain, resulting in a potent ability of these MMPs to bind to and cleave gelatin. In addition to gelatin, both gelatinases and stromelysins act to degrade major constituents of basement membranes, including type IV collagen, laminin, and fibronectin. The stromelysin enzymes (i.e., MMP-3, MMP-7, MMP-10, and MMP-11) exhibit activity toward a broad and diverse array of not only ECM substrates but also other MMPs, growth factors, and cytokines, such as insulin-like growth factor-binding proteins and tumor necrosis factor-
. The ability of these enzymes to cleave binding proteins and extracellular domains of growth factors expands the repertoire of MMP actions to include modulation of cell growth directly by controlling cell-matrix interactions or indirectly by regulating growth factor bioavailability. The MT-MMPs contain a transmembrane domain near their carboxy-terminal region that localizes these proteinases to the plasma membrane, and an extracellular domain directs a portion of the enzyme to the exterior surface of the cell. A major role of the MT-MMPs is the regulation of MMP-2 activation (discussed in detail below).
Tissue Inhibitors of Metalloproteinases
The activity of the MMPs in the extracellular space is rigorously controlled by MMP inhibitors, of which two major classes are generally distinguished: serum-borne and tissue-derived inhibitors [5, 6]. The serum-borne inhibitors include the macroglobulins, such as 2-macroglobulin and 1-inhibitor-3, which have the ability to inhibit a broad range of proteases. The second group of inhibitors are locally produced, specifically inhibit MMPs, and are referred to as tissue inhibitors of metalloproteinases (TIMPs). Although the two classes of inhibitors were originally distinguished as serum borne or tissue derived based upon the site of synthesis and action, reports that the serum-borne inhibitor 2-macroglobulin is produced in the ovary [7, 8], and that the tissue-derived inhibitors are present in serum, have blurred the nomenclature distinction between these two classes. The current review focuses on the TIMPs, because this class of inhibitor is highly abundant in reproductive tissues, hormonally regulated, locally produced, and postulated to coordinate numerous ovarian and uterine processes.
The first member of the TIMP family that was identified, TIMP-1, is a secreted glycoprotein (29 kDa) that binds to and inhibits active MMPs on a 1:1 stoichiometric basis. Inhibition of MMP action occurs through the interaction of the N-terminal domain of TIMP with the active site on the catalytic domain and the substrate-binding groove. Since the discovery of TIMP-1, other TIMPs have been described, including TIMP-2 (23 kDa), TIMP-3 (glycosylated, 21 kDa), and TIMP-4 (24 kDa). Specifically, TIMP-2 is secreted, differentially regulated from TIMP-1, and proposed to act selectively on different MMPs [5, 9, 10]. For example, TIMP-1 preferentially binds to the collagenases and MMP-9, whereas TIMP-2 has a high affinity for MMP-2. Unlike TIMP-1 or TIMP-2, TIMP-3 is bound to the ECM and has been suggested to act as an additional regulatory stop point for MMP action [11, 12]. Recently, TIMP-4 has been cloned, and the limited preliminary information regarding this inhibitor's substrate specificity and mode of action suggests that it has traits similar to those of TIMP-2 [5, 6, 13].
It is important to note that not all TIMP action is inhibitory of MMP function. For example, one action of TIMP-2 is its ability to facilitate the activation of MMP-2. This activation of MMP-2 is brought about by the ability of the C-terminal domain of TIMP-2 to bind to pro-MMP-2 to form a noninhibitory pro-MMP-2/TIMP-2 complex (Fig. 2). This complex can be recognized and bound by the extracellular domain of MT1-MMP. Once the complex is bound by MT1-MMP, adjacent, unoccupied MT1-MMPs cleave the pro-domain of MMP-2 and activate the gelatinase. Thus, TIMP-2, in conjunction with the MT-MMPs, acts to regulate the specific site of activation of other MMPs.
|
In addition to their ability to regulate MMP action, evidence is emerging that TIMPs can act to regulate growth. Support for this concept is based upon reports that TIMPs promote embryo growth and development [14], have erythroid-potentiating action [15–17], are antiangiogenic agents [18, 19], stimulate cell growth in a variety of tissues [20], influence apoptosis [21], and recruit quiescent cells into the cell cycle [22]. Therefore, TIMPs may act as autocrine/paracrine factors in reproductive processes involving cellular proliferation, differentiation, and neovascularization. The observation that a TIMP-1:green fluorescent protein chimera binds to the surface of MCF-7 breast carcinoma cells and is translocated to the nucleus hints at a possible nuclear role for TIMP-1 [23]. Supporting a multifunctional role in the ovary are the findings that a TIMP-1-like protein stimulates steroid production in the testis and ovary [24], that TIMP-1 stimulates granulosa cell estradiol production [25], and that TIMP-1-deficient mice have altered serum levels of estradiol and progesterone during the estrous cycle [26].
|
OVARY |
|---|
|
TOPABSTRACTINTRODUCTIONOVARYUTERUSSUMMARYREFERENCES |
|---|
Growth of the ovarian follicle requires extensive cellular proliferation and remodeling of the ECM as the follicle differentiates from a small, primordial follicle with a single layer of granulosa cells to a large, preovulatory graafian follicle [27, 28]. Folliculogenesis is characterized by proliferation of the granulosa cells, differentiation of the thecal cells from the ovarian stroma, and deposition of a basement membrane separating the theca from the avascular granulosa cells. Unlike the granulosa cells, the thecal layer is well vascularized and contains circumferential collagen bundles. With continued development, the granulosa cells secrete a mucopolysaccharide-rich fluid that coalesces to form the antral cavity in the secondary follicle. The resultant mature follicle in the primate is approximately 400-fold larger than the initial primordial follicle and rests in an extracellular environment of collagen, laminin, and fibronectin [27, 28].
Investigators have proposed that the extensive changes in the ovarian ECM during folliculogenesis are accomplished, in part, by the MMP system. Although a definitive causative effect has yet to be demonstrated, correlative data implicate the MMP system in follicular growth. Such data indicate that the localization and expression patterns of the MMP system change in concordance with the dynamic ovarian structural changes associated with follicular growth. For example, Bagavandoss [29] observed that the pattern of MMP-2 and MMP-9 immunolocalization was markedly increased at the latter stages of follicular development. During early follicular growth in the neonatal rat ovary, immunoreactive MMP-2 was present in the granulosa cells and the surface epithelium, whereas MMP-9 was absent. With induction of follicular growth by eCG, immunoreactive MMP-2 was observed in the granulosa, theca, and interstitium of preovulatory follicles, whereas MMP-9 was present in the thecal and interstitial tissue [29]. The increase in immunoreactive MMPs during the latter stages of follicular growth corresponds with increases in mRNA expression patterns following eCG administration. Cooke et al. [30] reported an increase in the mRNA expression for MMP-2 and MMP-9 within 12–24 h after eCG administration and an increase in mRNA levels for MMP-13 at 36–48 h after gonadotropin treatment. Corresponding to the changes in mRNA expression was an increase in gelatinolytic and collagenolytic activity at the latter stages of follicular growth [30]. In other species, such as the goat, collagenase activity has been noted to increase with increasing follicular size [31]. The changes in MMP expression during follicular growth are summarized in Table 2. Although little is known about the MMPs during follicular growth in the human, in other species, the MMPs are present within the follicle, are stimulated by the events associated with follicular development, and have localization and expression patterns that coincide with the changes in follicular remodeling of the granulosa and thecal cell layers. Of particular interest is the finding that in the sheep, an increase in MMP-2 and MMP-9 activity is associated with follicular atresia following hypophysectomy, suggesting that MMPs may play a role in atresia [32].
|
The changes in the TIMPs during early follicular growth have received limited attention, but the emerging evidence suggests that the alterations in TIMP-1 parallel the change in MMPs (Table 2). In the rat, an increase in TIMP-1 mRNA expression follows eCG administration [33], as also has been noted for the MMPs [30]. For the other TIMPs, eCG treatment in the rat results in a slight decline in TIMP-3 mRNA levels and no change in TIMP-2 mRNA [33]. The TIMPs are synthesized in the theca, stroma, interstitial tissue, and germinal epithelium during follicular development [34–36]. Immunoreactive TIMP-1 protein has been observed in the thecal cells, the interstitial blood vessels, and the surface epithelium of the rat [29], as well as in the granulosa and thecal layers of healthy follicles and the oocyte of the ewe [37]. As specific antibodies are developed for the various MMPs and TIMPs, the cellular localization in different species will be come apparent.
The observation that TIMP-1 expression increases concomitantly with the increase in MMP would appear paradoxical, in that one would anticipate a decrease in inhibitor as enzyme increased. Parallel regulation of MMPs and their inhibitors, however, has been observed and postulated to maintain proteolytic homeostasis and to provide localized control of ECM degradation [1–3]. A working hypothesis for the ovarian MMP system would encompass the precise coordination of MMPs and inhibitors to regulate the location and extent of follicular remodeling. Although the definitive role of the MMP system in follicular growth awaits discovery, it is apparent that the MMPs and TIMPs are in the appropriate cellular compartments and are associated with an increase in follicular growth in various species. Such observations have led Garcia et al. [31] to propose that the MMP system may regulate normal follicular maturation and atresia to achieve the appropriate number of ovulatory follicles. Alternatively, the concept that MMP inhibitors exist only to regulate ovarian proteolysis may be overly simplistic. Ovarian TIMPs may be multifunctional, as noted previously, acting as autocrine/paracrine factors in cellular proliferation, differentiation, neovascularization, or steroidogenesis during folliculogenesis.
Ovulation
Ovulation is a dynamic, orchestrated process set in motion by the LH surge and culminating in the breakdown of the follicular wall and extrusion of the oocyte [38]. The LH surge initiates and synchronizes a series of biochemical events that includes synthesis and secretion of prostaglandins, progesterone, cytokines, growth factors, and proteolytic enzymes, such as the plasmin/plasminogen-activator system and MMPs. Proteolysis in the ovulatory process has been postulated to degrade the apical follicular connective tissue, thereby facilitating oocyte release [38]. A body of morphological and biochemical reports provides support for a paramount role of the MMPs in follicular rupture. Morphologically, a dissolution of the granulosa cell basement membrane [39] and fragmentation of the collagenous matrix at the apex of the follicular wall [40] occur in numerous species. Concomitant with these morphological changes in the ovarian ECM, a decrease in ovarian and follicular collagen occurs after the LH surge [38, 41, 42], especially at the apex [38, 42, 43]. This decrease in follicular collagen is postulated to occur via the action of a cascade of proteolytic events. An increase in ovarian MMP mRNA and activity occurs before ovulation. The LH surge or administration of exogenous hCG to mimic LH action induces an increase in the mRNA for the collagenases (i.e., MMP-1 and MMP-13), the gelatinases, and MMP-19 (Table 2) in a temporal pattern consistent with breakdown of the follicular wall in the rat [44–47], mouse [48], ewe [49], and nonhuman primate [50]. These changes in mRNA are reflected by an increase in proteolytic enzyme activity. For example, in the rat, a four- to fivefold increase occurs in gelatinase [47] and collagenase activity [44] following LH/hCG stimulation, which can be blocked by inhibiting the endogenous LH surge with sodium pentobarbital [51]. In the human, gelatinolytic activity in the follicular fluid increases fourfold between Days 5 and 13 of the menstrual cycle, and an increase in the active form of MMP-2 occurs during the periovulatory period [52]. Similar increases in gelatinolytic activity for MMP-9 were reported in ovine follicular fluid [53]. Further evidence for an obligatory role of the MMPs in ovulation is forthcoming from experiments utilizing exogenous chemical MMP inhibitors, which result in an inhibition of oocyte release both in vitro [54, 55] and in vivo [45]. In toto, these observations suggest that follicular rupture requires focal degradation of the ECM controlled by ovarian MMPs.
In conjunction with the periovulatory increase in MMPs, LH or hCG stimulates an increase in TIMP mRNA expression and inhibitor activity. The LH surge results in an increase in TIMP-1 mRNA expression in the mouse [48, 56], rat [46, 57, 58], ovine [59], and nonhuman primate [50]. Associated with the increase in TIMP-1 mRNA levels is an increase in inhibitor activity in whole ovaries [57, 60], isolated granulosa cells [57, 61], and follicular fluid [53]. The inhibitor activity reflects TIMP-1 [53] or TIMP-like activity [53, 57, 60] as well as activity from the macroglobulin class of inhibitors [60]. For the other TIMPs, differences and/or conflicting reports on the periovulatory mRNA expression patterns are found (Table 2). For example, TIMP-2 mRNA has been reported to increase in bovine periovulatory follicles [62] and nonhuman primate granulosa cells [50] but to be unchanged in mouse [48, 56] and rat ovaries [63] as well as in ovine preovulatory follicles [64]. These differences may be species or tissue specific, or they may reflect the more sensitive detection of TIMP-2 mRNA by reverse transcription-polymerase chain reaction in the nonhuman primate granulosa cells versus Northern blot analysis as employed in the other studies. The conflicting reports that TIMP-3 mRNA increases during early proestrus in cycling mice [56] but is unchanged following hCG administration to eCG-primed immature mice [48] may be accounted for by differences in cycling versus gonadotropin-primed models, including responsiveness to hormonal signals or presence of the corpus luteum (CL).
Corpus Luteum
Following ovulation, the ruptured follicle is transformed into a CL by extensive cellular reorganization and neovascularization as the fluid-filled antral cavity is infiltrated by blood vessels, fibroblasts, and thecal and granulosa cells [65–67]. The degree of reorganization and intermixing of the thecal and granulosa cell layers during luteal formation differs among species. In nonprimate mammals, extensive cellular migration and reorganization occur. In primates, the intermixing of the different follicular cells is less extensive than in other species, such that a distinct granulosa-lutein and theca-lutein layer are formed [68]. With formation of the CL, progesterone production increases dramatically, yet negligible changes take place in the luteal connective tissue matrix during this period. At the end of the CL life span, two phases of luteolysis occur. First, a functional luteolysis is marked by a rapid decline in progesterone production. This is followed by a slower, prolonged structural luteolysis of the CL. The structural changes occur, in part, by an apoptotic mechanism involving proteolysis and phagocytosis [69–71]. Further degradation is accomplished by invading macrophages. Macrophages, which contain MMPs or can stimulate enzyme production in other tissues, increase in number in the involuting CL and phagocytize the luteal cells [69, 72].
The MMP system has been explored during all three phases of the luteal cycle: formation, maintenance, and regression (Table 2). During formation of the CL, the MMPs are elevated in the rat [73, 74], bovine [75], and human [76]. Using a pseudopregnant rat model, an increase in mRNA expression and activity for the gelatinases [73, 74] is found during luteal development. Interestingly, collagenolytic activity is elevated during luteal formation without concomitant changes in MMP-13 mRNA expression [73]. In the human, mRNA expression for MMP-9 was elevated during the early luteal period, whereas MMP-1 and MMP-2 levels were unchanged [76]. During the period of luteal maintenance, when progesterone production is maximal, mRNA expression patterns in the rat are at basal levels and unchanged for the gelatinases, MMP-13, and MT1-MMP [73, 74]. During luteal regression, however, a marked induction of MMP-13 mRNA but, surprisingly, no change in the mRNA expression of MMP-2, MMP-9, or MT1-MMP occurs in the rat [73, 74]. Although the mRNA for the gelatinases is unchanged during luteal regression in the rat, gelatinolytic activity is present. Endo et al. [77] observed an increase in MMP-2 activity during this period. In ovaries collected from mice throughout pregnancy, the temporal pattern of MMP-2 mRNA showed a marginal increase during gestation but a striking increase in the expression of MMP-10 mRNA at term [78]. In the human, MMP-2 and MMP-9 were elevated during the late luteal period, and administration of hCG to mimic luteal rescue in vivo was associated with a reduction in the expression and activity of MMP-2 [76]. Although species differences are apparent in the expression patterns of the various MMPs, the general model that is emerging is one in which MMPs are elevated during the period of extensive luteal ECM remodeling as the follicle is transformed into the CL. During the midluteal period, when the CL is fully formed and steroidogenesis is maximal, MMP activity is at basal levels. With the onset of luteal regression, the MMPs are again called into action for the remodeling and removal of the CL. Differences in the type and pattern of MMP expression among species may reflect differences in the intermixing of the different follicular cells to form the CL, differences in the composition of the luteal ECM, or physiological differences in the CL, such as the trophic influence of hCG versus placental lactogen, the absence or presence of progesterone receptors, or the local action of growth factors and cytokines.
Since the first description of TIMP-1 mRNA in the mouse CL [79], the changes in expression patterns throughout the luteal period have been explored in numerous species (Table 2). For TIMP-1, the mRNA expression is elevated during early luteal formation in the rat [74, 80] and bovine [62, 81]. In bovine granulosa cells undergoing luteinization in vitro, Zhao and Luck [82] noted that TIMP-1 mRNA is present and markedly increases during 4 days in culture. In contrast to the induction of TIMP-1 mRNA during luteal formation in the rat and bovine, TIMP-1 mRNA was unchanged throughout the luteal period in the ovine [59] or human [83], suggesting species-specific expression of this inhibitor. The TIMP-1 mRNA was highly expressed in large luteal cells compared to small luteal cells in the ewe [59] and was expressed in the granulosa-lutein cells of the human [83]. Immunolocalization of TIMP-1 protein supports the localization of mRNA to large luteal cells of the ewe [37, 84] or the granulosa-lutein cells of the human [83]. Of particular interest is that in the ewe, TIMP-1 was found in secretory granules in CL-containing oxytocin, and that TIMP-1 protein was seen undergoing exocytosis on Day 10 of the estrous cycle [84].
During the period of luteal maintenance, TIMP-1 and TIMP-2 mRNA expression in the rat does not change in whole-ovarian extracts, whereas TIMP-3 mRNA actually increases [80]. In CL isolated from ovine or bovine ovaries collected during the midluteal period, TIMP-2 mRNA was elevated [62, 64]. Smith et al. [64] demonstrated that TIMP-2 mRNA expression is abundant in large luteal cells and, by Western blot analysis, that these cells secrete TIMP-2.
During regression, TIMP-1 mRNA increases in the rat [74, 80], mouse [78], and bovine [85]. In contrast, TIMP-1 mRNA was unchanged in the human CL during the late luteal period [83] and decreased in the nonhuman primate CL following induction of luteolysis with either prostaglandin F2 or GnRH antagonist [86]. The temporal expression patterns for the other TIMPs indicate that the expression of TIMP-2 mRNA remains unchanged in the rat [80] and human [76]. In the mouse, ovarian TIMP-2 mRNA showed a marginal increase during gestation [78], whereas in the ovine and bovine, TIMP-2 mRNA declined to low levels during the late luteal period [62, 64]. To date, TIMP-3 has received limited attention, but in the rat, TIMP-3 mRNA expression has been found to be elevated at the end of pseudopregnancy [80]. The changes that occur in the MMP system follow the extensive ECM remodeling that occurs with luteal formation, maintenance, and regression. With the advent of specific MMP inhibitors and current gene-deletion models, the precise roles of the MMPs and TIMPs will begin to be elucidated.
|
UTERUS |
|---|
|
TOPABSTRACTINTRODUCTIONOVARYUTERUSSUMMARYREFERENCES |
|---|
The endometrium, which lines the largely muscular uterine body, is a highly specialized tissue that has evolved to support the dynamic events required to establish and maintain pregnancy. Structurally, the endometrium is composed of glandular and surface epithelium with a supporting stroma, which has a rich vascular supply. Ovarian steroid production directs a species-dependent pattern of endometrial growth and functional differentiation in preparation for pregnancy. In humans, the menstrual cycle may be repeated as many as 400 times during a reproductive life span [87]. In general, far less tissue remodeling occurs during the estrous cycle exhibited by many lower mammalian species compared to primates. Therefore, within the uterus of rodents, only moderate expression of MMPs has been observed compared to more robust MMP expression during the tissue-loss and -repair processes that are the hallmark of the menstrual cycle. Nevertheless, cyclic changes of MMP mRNA expression have been reported during the estrous cycle in rodents, with MMP-2, MMP-7, and MMP-11 being most abundantly expressed during estrus and proestrus, whereas MMP-3 and MMP-10 appear to be expressed to a much lesser extent [88, 89]. The expression of MMPs in the uterus of domestic species, including the cow [90] and sheep [91], has largely been reported only in association with the tissue dynamics of pregnancy as opposed to changes in uterine architecture throughout the estrous cycle. Given the more extensive involvement of MMPs in the tissue dynamics of the menstrual cycle compared to species exhibiting an estrous cycle [92], specific examples of endometrial MMP and TIMP expression discussed herein will draw principally from nonhuman primate and human studies.
The menstrual cycle consists of a series of highly co-ordinated, steroid-driven events: endometrial sloughing, tissue repair, and cellular proliferation and differentiation in preparation for pregnancy. In the absence of successful nidation, menstruation marks both the end of a failed cycle and the beginning of a new cycle. Re-epithelialization of the luminal endometrial surface is one of the first steps necessary for endometrial repair and is associated with expression of MMP-7, an epithelial-specific MMP [89, 93–95]. This MMP is also focally expressed by glandular epithelium as these tubular structures grow and elongate distally from the basalis region during the proliferative phase of the cycle [94, 95]. As opposed to expression of only MMP-7 by endometrial epithelial cells, stromal cells express numerous MMPs during the proliferative and menstrual phases (Table 3). In the human endometrium, a cyclic, but variable, expression pattern of specific mRNA for MMP-1, MMP-2, MMP-3, MMP-9, MMP-10, and MMP-11 has been detected in the stroma during the menstrual cycle [96–98].
|
A detailed description of endometrial MMP gene regulation is beyond the scope of this review, but it is important to note that in response to ovarian steroids, a variety of growth factors and cytokines are expressed in the endometrium [99, 100] that can impact MMP expression. Complex mechanisms, mediated by steroids and local tissue-specific factors, act in concert to promote the selective expression of specific MMPs while limiting the expression or action of others [92, 101]. Additionally, structural components of the ECM can interact with specific cellular receptors, such as members of the integrin family, to impact MMP expression during these periods of tissue repair and remodeling [102–105]. Although it is uncertain whether estrogen can directly affect the cell type-specific expression of MMPs in the uterus of any species during periods of growth [95, 106, 107], progesterone exposure and subsequent withdrawal profoundly impacts the expression patterns of multiple endometrial MMPs [94]. In vitro studies by multiple laboratories have shown that progesterone is a potent inhibitor of endometrial MMP mRNA expression and protein secretion [89, 106, 108–114]. Progesterone clearly acts to suppress endometrial MMP expression, but the mechanism(s) by which progesterone regulates transcription of specific MMP genes is not understood. The promoter regions of MMPs that are suppressed by progesterone in vitro lack traditional progesterone-response elements; thus, regulation of certain MMP genes by this steroid may occur via nonclassical DNA sequences [92]. Additionally, cell-cell communication appears to be a necessary component in the regulation of cell-specific MMP expression following steroid exposure. For example, isolated human endometrial epithelial cells require the presence of the stroma as well as the action of transforming growth factor-ß for suppression of MMP-7 protein secretion in response to progesterone [106]. In contrast, in isolated stromal cells, progesterone treatment alone suppresses MMP-3 expression and prevents the subsequent stimulation of this MMP by inflammatory cytokines [115]. These and other data support a critical role for progesterone in the suppression of endometrial MMPs in humans, but paracrine factors also may be important for MMP expression. For example, Rudolph-Owen et al. [95] demonstrated in the nonhuman primate that growth-related MMP expression disappears completely at midcycle, before the postovulatory rise in progesterone.
As noted above in relation to ovarian tissues, the activity of secreted and activated MMPs can be further regulated within uterine tissues by TIMPs. Rodgers et al. [94] were among the first to identify TIMP-1 mRNA in both endometrial epithelial and stromal cells with varying intensity of expression throughout the menstrual cycle. The TIMP-1 message in epithelium appeared low regardless of the cycle phase, although stromal cell expression of TIMP-1 mRNA progressively increased through the proliferative and secretory phases (Table 3), with the greatest expression in menstrual tissues [94]. Additionally, the expression of TIMP-1 in small arteriolar and capillary vascular tissue in secretory endometrium suggests that TIMP-1 may serve to promote vascular integrity during the invasive events of implantation [92, 98]. Immunoreactive TIMP-1, TIMP-2, and TIMP-3 proteins have each been reported within various tissue compartments of the human endometrium [116]. The lowest levels of TIMP-1 protein were observed during the mid to late proliferative phase, compared to more intense expression during the late secretory phase, associated mostly with epithelial cells. Compared to the expression of TIMP-1, that of TIMP-2 and TIMP-3 in human endometrium (Table 3) appears to be less intense among multiple cell types, with only moderate variation throughout the menstrual cycle [116].
Menstruation
At the end of each menstrual cycle, declining steroid support leads to extensive endometrial tissue loss in humans and other nonhuman primates. Both estrogen and progesterone likely are necessary for maintaining overall tissue integrity in the endometrium, because organ cultures of endometrial tissue maintained in the absence of these steroids release large amounts of numerous MMPs that can be quickly activated as measured by zymography [110]. Not surprisingly, the broadest and most extensive in vivo MMP expression in the endometrium (Table 3) occurs at menstruation [94, 98]. The role of progesterone in stabilizing the endometrium may involve both preventing MMP expression and blocking the action of inflammatory cytokines [110, 115]. For example, physiologic concentrations of progesterone (10–200 nM) almost totally abolish MMP-1 release as well as MMP-2 and MMP-9 activity in explant cultures of human endometrium [109]. In contrast, withdrawal of progesterone, either in vivo or in vitro, results in the rapid expression of multiple MMPs, which is initiated, at least partially, through the localized actions of inflammatory cytokines [109, 110, 115, 117, 118]. In an isolated cell culture model designed to mimic menstruation, Lockwood et al. [109] demonstrated that MMP-1 mRNA expression was significantly increased in isolated stromal cells following 4 days of treatment with mifepristone (RU-486), which blocks progesterone action, compared to cultures maintained with estradiol and progesterone. Another detailed survey of multiple MMPs demonstrated increased secretion of MMP-1, MMP-2, MMP-3, and MMP-9 by isolated stromal cells following progesterone withdrawal [118]. Although an increased expression of TIMP-1 has been noted in menstrual tissue [98], no effect of progesterone withdrawal has been noted on either TIMP-1 or TIMP-2 mRNA expression or protein secretion in the in vitro model of menstruation [118].
Early Pregnancy
In contrast to the very low levels of MMP expression among animals lacking the tissue loss of menstruation, the dynamics of implantation, placentation, and parturition requires the participation of numerous MMPs in a broad range of mammalian species. The MMPs are responsible for the basement membrane and ECM degradation required for the invasive processes necessary to establish pregnancy (Table 3). The invasion of the endometrium by trophoblast cells during the process of implantation and placentation involves dramatic MMP-mediated cellular mobility, host-tissue remodeling, and placental development. Early reports noted that cytotrophoblasts isolated from first-trimester human placenta degraded basement-membrane substrates in vitro, whereas no degradative activity was observed in second- or third-trimester cytotrophoblast cells [119, 120]. The various trophoblast phenotypes found in humans and Old World primates were soon shown to express mRNA and positive immunoreactivity for multiple MMPs and TIMPs, including MMP-2, MMP-3, MMP-9, MMP-11, MT1-MMP [121–131], and TIMP-1 and TIMP-3 [132, 133]. Among other species, such as the rodent and domestic species, MMP-1, MMP-2, MMP-3, MMP-9, TIMP-1, and TIMP-2 each appear to contribute to endometrial remodeling at the time of placental development [91, 134, 135]. Interestingly, the expression of MMP-7 may be more prominent in rodents [136] during the establishment of pregnancy than in humans [130]. In contrast, MMP-7 expression is very prominent at ectopic sites of pregnancy in humans [137]. Among the various MMPs expressed during the establishment of pregnancy, it appears that MMP-9 is the primary MMP secreted by trophoblast cells to facilitate endometrial invasion [138]. Shimonovitz et al. [139] found that primary cultures of trophoblast cells isolated from first-trimester pregnancy constitutively express both MMP-2 and MMP-9 on a gelatin matrix, with higher levels of MMP-9 expression and activity compared to those in cells from the third trimester.
Maintaining tissue integrity during the disruptive events of implantation and placentation is critical for normal uterine function and successful reproduction. The endometrium, in response to estrogen and progesterone, develops a unique ability to limit local tissue MMP expression in the face of continued proteolytic activity by the invading trophoblasts. However, investigators have demonstrated that treatment of trophoblast cells with progesterone also decreases steady-state levels of MMP-9 mRNA in a concentration-dependent manner. Suppression of both MMP-9 message and gelatinase activity could be reversed by addition of the progesterone-receptor antagonist onapristone [140]. The production of progesterone by the ovary and, subsequently, by the placenta in humans may be critical to maintaining tissue stability at the maternal-fetal interface, because stromal and decidual cells can express MMP-1, MMP-2, MMP-3, MMP-9, and MT1-MMP [115, 121–124, 127, 130]. Interestingly, it was recently demonstrated that progesterone exposure of endometrial stromal cells, either in vivo or in vitro, dramatically reduces the ability of interleukin (IL)-1 to stimulate MMP-3 expression [115]. Cytokine involvement in maternal-fetal communication was first implicated when IL-1 mRNA and protein were detected in human cyto- and villous trophoblasts and IL-1 and IL-1ß were reported to be produced by blastocysts in vitro [141–143]. Members of the IL-1 family are produced by the invading trophoblasts [142] and are thought to participate in the establishment of pregnancy, including the regulation of MMP expression. However, progesterone is such a potent inhibitor of endometrial MMP expression that a loss of progesterone receptors at the implant site has been suggested to be necessary for successful implantation in numerous species, including humans [144–147].
|
SUMMARY |
|---|
|
TOPABSTRACTINTRODUCTIONOVARYUTERUSSUMMARYREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 Supported by grants to T.E.C. (NIH HD23195, HD34400, and the Lalor Foundation) and to K.G.O. through cooperative agreement U54 HD37321 as part of the Specialized Cooperative Centers Program in Reproductive Research and by the Endometriosis Association Research Program at Vanderbilt University. 
2 Correspondence: Thomas E. Curry, Jr., Department of Obstetrics and Gynecology, University of Kentucky Medical Center, 800 Rose Street, Room C355, Lexington, KY 40536-0293. FAX: 859 323 1931; tecurry{at}pop.uky.edu
Accepted: November 29, 2000.
Received: September 19, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONOVARYUTERUSSUMMARYREFERENCES |
|---|
2-macroglobulin in rat ovarian follicles and corpora lutea. Endocrinology 1989; 125:2985–2995[Abstract]2-macroglobulin gene in rat ovarian granulosa cells: stat 5 activation and binding to the interleukin-6 response element. Mol Endocrinol 1997; 10:171–184[Abstract]-macroglobulins in the ovulating rat ovary: possible regulators of collagen matrix breakdown. Biol Reprod 1991; 45:334–342[Abstract]
, a decrease in expression is shown as 
, no change in expression is depicted as –, and unknown is represented by ?