This Article Abstract Full Text (PDF) All Versions of this Article: 27/1/17    most recent Final Manuscript Author Manuscript Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Jabbour, H. N. Articles by Critchley, H. O. D. Search for Related Content PubMed PubMed Citation Articles by Jabbour, H. N. Articles by Critchley, H. O. D.

Endocrine Regulation of Menstruation

Medical Research Council Human Reproductive Sciences Unit (H.N.J., R.W.K., H.M.F.), and Reproductive and Developmental Sciences (H.O.D.C.), University of Edinburgh, Centre for Reproductive Biology, The Queen’s Medical Research Institute, Edinburgh EH16 4TJ, United Kingdom

Correspondence: Address all correspondence and requests for reprints to: Henry N. Jabbour, Medical Research Council Human Reproductive Sciences Unit, The Queen’s Medical Research Institute, 47 Little France Crescent, Edinburgh EH16 4TJ, United Kingdom. E-mail: h.jabbour{at}hrsu.mrc.ac.uk

	Abstract

Top Abstract I. Introduction II. Endometrial Morphology III. Steroid Control in... IV. Progesterone Withdrawal and... V. Repair and Vessel... VI. Disorders of Menstruation VII. Local Mediators Associated... VIII. Perspective and Future... References

 
In women, endometrial morphology and function undergo characteristic changes every menstrual cycle. These changes are crucial for perpetuation of the species and are orchestrated to prepare the endometrium for implantation of a conceptus. In the absence of pregnancy, the human endometrium is sloughed off at menstruation over a period of a few days. Tissue repair, growth, angiogenesis, differentiation, and receptivity ensue to prepare the endometrium for implantation in the next cycle. Ovarian sex steroids through interaction with different cognate nuclear receptors regulate the expression of a cascade of local factors within the endometrium that act in an autocrine/paracrine and even intracrine manner. Such interactions initiate complex events within the endometrium that are crucial for implantation and, in the absence thereof, normal menstruation. A clearer understanding of regulation of normal endometrial function will provide an insight into causes of menstrual dysfunction such as menorrhagia (heavy menstrual bleeding) and dysmenorrhea (painful periods). The molecular pathways that precipitate these pathologies remain largely undefined. Future research efforts to provide greater insight into these pathways will lead to the development of novel drugs that would target identified aberrations in expression and/or of local uterine factors that are crucial for normal endometrial function.

I. Introduction

II. Endometrial Morphology

A. Endometrial structure

B. Structure of vessels

C. Endometrial leukocyte populations

III. Steroid Control in Endometrium

A. Endometrial steroid receptor expression

B. Endometrial paracrinology

C. Endometrial intracrinology

D. Progesterone and progesterone receptors

E. Estrogen and estrogen receptors

F. Androgens and the androgen receptor

G. Glucocorticoids and the glucocorticoid receptor

IV. Progesterone Withdrawal and the Mechanisms of Menstrual Bleeding

A. Progesterone action in the endometrium

B. Decidual changes preceding menstruation

C. Physiological withdrawal of progesterone

D. Menstruation, the sequence

E. Origin and control of MMPs—the final effectors

F. Animal models of menstruation

V. Repair and Vessel Regrowth

A. Epithelial growth

B. Pattern of angiogenesis during the cycle

C. Angiogenic factors and their receptors in the uterus

D. Localization and changes in angiogenic factors throughout the cycle

E. Regulation of endometrial angiogenic factors

F. Effects of manipulation of angiogenic factors: in vivo models

VI. Disorders of Menstruation

A. Parameters of normal menstruation

B. Menstrual dysfunction

C. Menorrhagia

D. Dysmenorrhea

VII. Local Mediators Associated with Menstrual Dysfunction

A. Prostanoids

B. Other angiogenic/permeability factors

VIII. Perspective and Future Direction

	I. Introduction

Top Abstract I. Introduction II. Endometrial Morphology III. Steroid Control in... IV. Progesterone Withdrawal and... V. Repair and Vessel... VI. Disorders of Menstruation VII. Local Mediators Associated... VIII. Perspective and Future... References

 
THE UTERUS PLAYS a crucial role in survival of the species in all viviparous animals. Implantation of the fertilized egg is a critical event common to all species, but humans and old-world primates differ from most other animals in that they shed a significant proportion of their endometrium if pregnancy is not established at the opportune time. Human endometrium is thus a dynamic tissue that, to prepare for implantation, undergoes well-defined cycles of proliferation, differentiation, and shedding (menstruation) in response to the prevailing endocrine and paracrine environment. However, the process of menstruation may be accompanied by distressing symptoms such as menorrhagia (excessive menstrual blood loss) and dysmenorrhea (painful periods) that are still little understood.

Regular menstrual bleeding is the outward manifestation of cyclical ovarian function. The average woman today in developed countries may expect to menstruate over 400 times during her reproductive life span. In contrast, in less well-developed countries and before the availability of reliable contraception, which has allowed women to regulate their own fertility, the majority of women were amenorrheic for most of their lives. This feature was due to later puberty, high numbers of pregnancies, and prolonged lactation.

It is essential to have a detailed knowledge of the mechanisms regulating endometrial events involved in implantation and menstruation if we are to understand the mechanisms responsible for abnormal menstrual bleeding, early pregnancy failure, and infertility. Indeed, only with a better understanding of the local mechanisms involved in endometrial function will progress be made to modulate sex steroid interactions in target cells. This review will focus on endocrine and paracrine regulation of menstruation and the local molecular aberrations associated with menstrual dysfunction.

	II. Endometrial Morphology

Top Abstract I. Introduction II. Endometrial Morphology III. Steroid Control in... IV. Progesterone Withdrawal and... V. Repair and Vessel... VI. Disorders of Menstruation VII. Local Mediators Associated... VIII. Perspective and Future... References

 
The human endometrium is a dynamic tissue that, in response to the prevailing steroid environment of sequential ovarian estrogen and progesterone exposure, undergoes well-characterized cycles of proliferation, differentiation, and tissue breakdown on a monthly basis. If pregnancy fails to be established, then the endometrium is shed and regenerates. Menstruation is the reproductive process whereby the upper two thirds of the endometrium (functional layer) is shed and regenerated on a repetitive basis. The endometrium is consequently a site of recurrent physiological injury and repair (1).

Studies undertaken by Markee (2) and Corner and Allen (3) established the role for ovarian steroids, estradiol, and progesterone in regulating the changes in endometrial conformation across the menstrual cycle. Progesterone is essential for the establishment and maintenance of pregnancy consequent upon the transformation of an estrogen-primed endometrium. Sex steroids, acting via their cognate receptors, initiate a cascade of gene expression and events crucial for successful implantation and early stages of pregnancy. Application of knowledge from the human genome, utilizing microarray technologies, has allowed several groups to contribute to a rapidly expanding literature on gene profiles during the "putative window" of implantation (4, 5, 6).

Menstruation is the response of the endometrium to the withdrawal of progesterone (and estrogen) that occurs with the demise of the corpus luteum in the absence of pregnancy (Fig. 1). The molecular mechanisms by which sex steroids induce these events within the endometrium at the time of menstruation, involves complex interactions between the endocrine and immune system (1). Crucial structural components in the endometrium during the menstrual process are the component blood vessels and the dynamic population of leukocytes that influx at this time (Fig. 2). Only with a better understanding of the local mechanisms involved in the regulation of endometrial structure and function will there be the potential to pharmacologically modulate endometrial function in a way that would offer new strategies in the management of female reproductive health.

View larger version (47K): [in this window] [in a new window]   FIG. 1. Illustration of the alternatives for a progesterone-primed endometrium. The progesterone induced changes that characterize the endometrium primed for implantation also bring about progesterone dependency of the tissue. On the demise of the corpus luteum, falling progesterone levels initiate an inflammatory process that starts in the stromal compartment but involves leukocyte immigration and metalloproteinase (MMP) activation. Eventual shedding of the functional layer of the endometrium opens the way for estrogen-dependent regrowth of the tissue.

View larger version (147K): [in this window] [in a new window]   FIG. 2. Distribution of CD56+ve uNK cells in endometrium during the secretory phase of the cycle. This shows the close clustering of the uNK cells (solid arrow). The spiral vessels are indicated by open arrows.

The progesterone-dominated latter half of the menstrual cycle is constituted by an early, mid, and late secretory phase. The pattern of sex steroid receptor expression in the endometrium across the secretory phase reflects the fact that the early secretory phase is regulated by both estrogen and progesterone; the mid secretory phase is regulated by progesterone alone as estrogen receptor (ER) is down-regulated in the glands and stroma at this time (9); and the late secretory phase is associated with progesterone withdrawal and, consequently, menstruation. Evidence that endometrial genes are regulated has been demonstrated by the observed changes in gene expression from late proliferative to mid secretory (5) and early to mid secretory phase (4) in microarray studies. Interestingly, the failure to make a transition in gene expression has been demonstrated in endometriosis (10) with dysregulation of specific genes during the mid secretory phase in this condition.

It is notable that the exogenous administration of sex steroids produces a marked modulation of the classic histological features such as the glandular structure, mitotic status of glandular cells, and secretions in the lumen of the glands (7) when compared with accurately dated endometrium collected during a physiological cycle (11). Description of the morphological features of the endometrium may be related to the timing of ovulation. It has become widely accepted that when histological dating is in excess of 2 d ahead of classical anticipated histological features, the endometrium is advanced; if greater than 2 d delayed, then it is histologically described as delayed (12). Much controversy over endometrial dating exists (13, 14). It is likely that better methods for evaluating endometrial dating are now available and that a consistency across a set of parameters, such as date of last menstrual period in the context of regular cycles, histological dating, and serum endocrine profile on the day of biopsy, are important for accurate dating. Precision can also be obtained if chronological dating is based on determination of the LH surge (15) or timing of ovulation as defined with pelvic ultrasound (16). A morphometric analysis that provides a quantitative description of endometrial structure may also be used (12, 17). Li and Cooke (12) described five measurements, from a total of 17 morphometric parameters that are necessary to achieve a reliable correlation. These features are the volume fraction of the gland occupied by glandular cells, predecidual reaction, luminal secretion, pseudo stratification, and glandular mitoses.

Recently, the utility of histological dating of endometrium in the evaluation of infertile couples has been questioned. Coutifaris et al. (13) have reported on a prospective multicenter study where the out-of-phase biopsy results failed to provide good discrimination between fertile and infertile couples. Furthermore, a randomized observational study by Murray et al. (14) using histological features identified by both objective and systematic analyses failed to reliably distinguish a specific menstrual cycle day or narrow interval of days. This latter study concluded that histological dating had neither the accuracy nor the precision to provide a guide for clinical management. Additional criteria may emerge following studies such as endometrial stage prediction based on global gene expression (18). In this study, analysis was made of some 10,000 genes in endometrium, collected from across the cycle and histologically divided into seven stages: early, mid, and late proliferative; early, mid, and late secretory; and menstrual, according to criteria of Noyes et al. (7).

Classic histological criteria may consequently be inappropriate for the assessment of endometrium exposed to exogenous steroids, for example progestogens; or novel steroid receptor modulators. Indeed, it is likely that the endometrial features exhibited after exposure to steroid receptors modulators will require new descriptors to aid interpretation of possible unique and previously unreported effects upon endometrial cellular components. The endometrial features displayed in an endometrial biopsy will be indicative of the timing in the cycle of administration, route of steroid delivery and formulation, dose of compound administered, and the duration of therapy (19, 20, 21).

B. Structure of vessels
The blood vessels of the endometrium are critical to menstruation. The spiral form of the arterioles in the upper two thirds of the functional layer (2) are characteristic of menstruating species. Such vessels are involved in both leukocyte entry and vasoconstriction and clearly involved in menstruation. The strength of the small arterial blood vessels in endometrium is derived from a combination of endothelial cells, basement membrane, and cells with smooth muscle character that surround this membrane. By the late secretory phase, the spiral arterioles are surrounded by a characteristic cuff of cells that resemble the decidual cells of pregnancy (22). These cells possess smooth muscle character (e.g., actin expression) as do all decidual cells (23). However, despite the similarities between the perivascular and decidual cell, a distinction can be clearly seen when progesterone is withdrawn from decidua. Immunohistochemical studies show the distinctive cuff, expressing inflammatory agents such as prostaglandins and cytokines, clearly defined against a background of decidual cells (24, 25). These perivascular cells appear to differ from the pericyte (the smooth-muscle cell that wraps around normal blood vessels) because they form a thicker layer, but they may still have processes that reach through the basement membrane and interact with endothelial cells.

The basement membrane, comprised of collagen type 4, fibronectin, and glycoseaminoglycans, varies from 50 to 350 nm in thickness, with an increase occurring during the luteal phase of the cycle. Certain components such as heparin sulfate proteoglycan show a decrease in the menstrual phase of the cycle (26) which may reflect destabilization of the vessels. The basement membrane can be broken down by various matrix metalloproteinases (MMPs) and it is the control of these MMPs by steroid hormones that is one conduit for progesterone action. The matrix components of the basement membrane are synthesized by neighboring (perivascular) stromal cells under the influence of progesterone. In particular, progesterone stimulates synthesis of fibronectin (27) and thrombospondin (28). Tenascin expression appears to reflect areas of proliferating cells (29) and thus, although global levels fall during the secretory phase of the cycle, they increase in cells immediately surrounding the blood vessels (30) where there is also an increase in proliferation markers such as Ki67 (31). Moreover, studies in the rat suggest that tenascin is clearly evident in the mesometrial gland region where it may interact with the uterine natural killer (uNK) cells in that area. These findings underline the potential for stromal cell–uNK cell interactions in women.

Some components of the membrane such as heparin sulfate proteoglycan decrease during the menstrual phase of the cycle (26). Because of the overall negative charge on many components of the extracellular matrix, these molecules can act as tethering points for growth factors. Thus, loss of molecules such as heparin sulfate proteoglycan during the menstrual process may lead to an increase in growth factor availability that will induce the regrowth of endometrial tissue (26).

C. Endometrial leukocyte populations
A dynamic leukocyte population exists within the endometrial stroma, the numbers and types of which vary across the menstrual cycle and throughout pregnancy. Endometrial leukocytes include T and B cells, mast cells, macrophages, and neutrophils. It is the phenotypically unique uNK cells that make up the majority of the leukocyte population in the late secretory phase and early pregnancy (32, 33). uNK cells are the major leukocyte population present in the endometrial stroma at the time when implantation, placentation, and decidualization occur. In the absence of pregnancy, uNK cells may be important in the initiation of menstruation. The observed cyclical increases in uNK cell numbers in the endometrium implicate direct or indirect regulation by endocrine signals, these being estrogen and/or progesterone. uNK cells have a unique phenotype (CD56 bright, CD16-, CD3-), which distinguishes them from peripheral blood NK cells (CD56 dim, CD16 bright, CD3-).

In the proliferative phase, few uNK cells are apparent but their numbers increase from day LH+3 and particularly so in the mid to late secretory phase (day LH+11–13) where they are located in close contact with endometrial glands and spiral blood vessels (32, 34). It remains to be established whether the increase in cell number is solely the result of in situ proliferation or whether there is also de novo migration from the peripheral circulation. A precursor cell type might be selectively recruited into the endometrium, where it differentiates to become the uterine-specific NK cell. In support of this theory is the existence of a subset of peripheral NK cells (around 1% of total circulating NK cells) that express a similar antigenic phenotype to uNK cells (35). On the other hand, proliferation of uNK cells in the endometrium has been described using the proliferation marker Ki67 (36, 37).

Quantitative real time RT-PCR studies have demonstrated an absence of ER and progesterone receptor (PR) mRNA in purified uNK cells (33). In contrast, mRNA for ERß isoforms (ERßcx/ß2, ERß1) and the glucocorticoid receptor (GR) have been localized in these cells (33). Colocalization using specific monoclonal antibodies has confirmed that uNK cells are immunonegative for ER and PR protein (33, 38, 39). uNK cells are also immunonegative for ERßcx/ß2 but do express ERß1 and GR proteins. These recent data have raised the possibility that estrogens and glucocorticoids could be acting directly on uNK cells through ERß and GR, respectively, to influence gene transcription in the endometrium and decidua (33).

Estrogen and progesterone may exert their effects on uNK cells indirectly via cytokines such as IL-15 and prolactin or other soluble factors (40, 41, 42). Because these factors are mainly secreted by the uterine stromal cell (USC) and because it is this cell that maintains PR, this is the likely conduit for progesterone action on uNK cells. Indeed, uNK cells and the endometrial stromal cell may have a special relationship as King (32) has commented that ectopic decidua is always associated with NK cells. Moreover, class I human leukocyte antigen (HLA; specifically HLA-B7) mRNA was increased on the surface of endometrial stromal cells when these cells were decidualized (43). HLA would interact with the killer inhibitory receptors on the NK cells and prevent lysis of the stromal cell by the NK cell. Additional interactions between the uNK cell and the stromal cell may involve prolactin. Prolactin is a 23-kDa neuroendocrine pituitary hormone that is also secreted by other cells such as the decidualized stromal cell (44). Prolactin stimulates proliferation of lymphocytes by stimulating IL-2 (45) and may have growth-promoting effects on other cells (46). Prolactin secretion may be under the influence of cytokines produced by hematopoietic cells, which include the uNK cells (47). Because the uNK cells have the prolactin receptor (42), a two-way interaction between stromal and uNK cell may contribute to homeostasis with first-trimester decidua. Although the uNK cells have no conventional PRs, it is possible that they respond to progesterone via nongenomic membrane receptors. Membrane receptors for progesterone have been reported in fish, and human counterparts of the , ß, and forms have been identified (48). However, although mRNA for these proteins has been found in several cell types, including endometrial cells (O. Harding, H. O. D. Critchley, H. N. Jabbour, R. W. Kelly, and T. A. Bramley, unpublished observations), there have been no confirming reports that these molecules play any role in progesterone signaling.

The processes of endometrial differentiation, menstruation, and placentation involve the remodeling of the endometrial vasculature. The angiogenic factor, vascular endothelial growth factor (VEGF)-A plays an important role in regulation of vascular permeability and the establishment of new blood vessel formation and induces endothelial cell proliferation, migration, and differentiation in the endometrium (49). Of the VEGF family, VEGF-A has also been localized to individual cells, presumed to be leukocytes, distributed throughout the endometrial stroma. These cells have been identified as neutrophils through dual immunohistochemical staining by Mueller et al. (50). VEGF-A expression has also been reported in uterine macrophages in the secretory phase of the cycle (51). VEGF-C and other angiogenic factors, placental growth factor, and angiopoietin (Ang)-2 mRNA are expressed in uNK cells (52). VEGF-C was originally characterized as a growth factor for lymphatic vessels, but it can also stimulate endothelial cell proliferation and migration (53). These patterns of angiogenic growth factor expression and the intimate spatial association of uNK cells with spiral arterioles implicate a role for these cells in endometrial angiogenesis.

There are few data pertaining to mechanisms of neutrophil recruitment into the human uterus. Small numbers are present in endometrium during the majority of the menstrual cycle, except immediately premenstrually and during menses (54). The withdrawal of progesterone in the late secretory phase of the cycle may be the trigger for neutrophil influx because in the sheep, the withdrawal of progesterone results in a rapid influx of polymorphonuclear leukocytes into the uterus (55). Neutrophils synthesize and release a wide range of immunoregulatory cytokines and thereby initiate and augment cellular and humoral immune responses. One recent observation relating to the role for neutrophils in mucosal defense as well as in the mechanism of menstruation is the expression in human endometrial neutrophils of the antiproteinase and antimicrobial molecule elafin in a menstruation-dependent manner (56).

Macrophages contribute approximately 20% of the total leukocyte population in late secretory (premenstrual) endometrium (57, 58). These cells are present throughout the menstrual phase but increase in number in the mid to late secretory phase and in decidua (59). Although macrophages show a cyclical pattern of expression across the cycle, they do not express classic ER or PR (38, 39). Control of their appearance by the ovarian sex steroids is, therefore, likely to be indirect. Endometrial macrophages display phenotypic differences, and subtypes have been described that express MMP-9 (60) and the membrane-bound MMP, MT1-MMP (61). Both these enzymes are involved in the breakdown of extracellular matrix and have been proposed to play a role in menstruation. Macrophages also produce a wide variety of regulatory molecules (58) that could stimulate the production of MMPs and proinflammatory cytokines from adjacent cells. Furthermore, macrophages are also a source of VEGF (62). Production of VEGF and the up-regulation of macrophage numbers premenstrually may implicate these cells in the menstrual process, where hypoxia and VEGF could lead to an induction of MMPs (63), and also in the revascularization of the endometrium after menstruation.

Another uterine leukocyte, the mast cell, has been identified in the endometrium in small numbers, mainly in the stromal compartment (64, 65). Human mast cells are hemopoietic cells that are characterized by their content of neutral protease and contain either tryptase alone or tryptase and chymase (66). Mast cell distribution and numbers are not altered during the menstrual cycle, but the cells are activated before menstruation when a diffuse pattern of immunoreactivity is observed. In vitro data have implicated a role for mast cells in the up-regulation of MMPs before the onset of menstruation (67, 68). Endometrial mast cells do coexpress the enzyme mast cell tryptase and MMP-1 in the same granules (69). Human endometrial mast cells do not express the genomic PR (H. Critchley, S. Milne, and S. Brechin, unpublished observation) and so progesterone is unlikely to be acting directly to regulate mast cell traffic in the endometrium across the cycle. Progesterone does, however, regulate MMP expression (including MMP-1) by other cell types that in turn may influence mast cell activation (67, 70).

	III. Steroid Control in Endometrium

Top Abstract I. Introduction II. Endometrial Morphology III. Steroid Control in... IV. Progesterone Withdrawal and... V. Repair and Vessel... VI. Disorders of Menstruation VII. Local Mediators Associated... VIII. Perspective and Future... References

 
A. Endometrial steroid receptor expression
Steroids interact with their target organs via specific nuclear receptors. Members of the nuclear receptor superfamily include PR, ER, GR, and androgen receptors (ARs). At least two isoforms of the human PR have been described (71, 72). PRA and PRB derive from a single gene and function as interactive transcriptional regulators of progestin-responsive genes. Two structurally related subtypes of ER, known as alpha (ER) and beta (ERß) have now been identified in the human and are derived from separate genes (73, 74, 75). The actions of steroids may also involve membrane as well as nuclear receptors. Some actions of estrogens are mediated via intracellular second messengers or other signal transduction pathways through nongenomic action (76). Recently, nongenomic (membrane bound) PRs have been characterized with no sequence homology to the nuclear receptor (48, 77).

The nuclear steroid receptors share common structure and functional domains, denoted A/B, C, D, E, and F (78). The A/B region is located at the N-terminal end and is not well conserved. This region (A/B) contains a transactivation domain (AF1). The C domain contains a highly-conserved DNA-binding domain consisting of two "zinc" fingers. Sequences within the C domain determine the specificity of the different receptors for specific hormone response elements. Aberrations in DNA and mutations in this region can result in receptor dysfunction. Next to the DNA-binding domain is the variable hinge region (D). The ligand-binding domain (LBD), region E, has a dimerization region and two transactivation domains (AF-2 and AF-2a). The LBD determines whether or not the receptor is activated. There are also accessory proteins involved in the stabilization or destabilization of the transcriptional complex necessary for steroid hormone action (79, 80). Further detailed discussion of the mechanisms of steroid receptor function lies beyond the scope of this review. However, it is worth noting that although most genes that respond to progesterone can be identified by progesterone response elements in their promoter region, this is not invariably so. A recent review highlights mechanisms by which progesterone sensitizes key kinase cascades to growth factors such as epidermal growth factor (EGF) and deals with synergistic up-regulation of growth genes such cyclin-D1 (81). Thus, although progesterone may restrict estradiol-driven endometrial growth in general, there are areas of growth in the progesterone-dominated uterus: for example, in cells surrounding the spiral arterioles (31). Moreover, the growth rate of endometrial stromal cells in culture is enhanced by a combination of progesterone and growth factors such as basic fibroblast growth factor (FGF) (82).

Estrogen is the steroid responsible for endometrial proliferation. Progesterone (and progestogens) will only result in differentiation if PRs are present in endometrial cells. PR expression requires previous exposure to estrogen. Progestogens exert an antiestrogenic effect with inhibition of endometrial growth and induction of maturation and differentiation of the glandular and stromal cells.

The expression of endometrial sex steroid receptors (PR, ER and ß, AR) varies temporally and spatially across the menstrual cycle (1, 9, 83, 84, 85). The expression of ER and PR is under dual control by estradiol and progesterone. Both endometrial ER and PR are up-regulated during the proliferative phase by ovarian estradiol and subsequently down-regulated in the secretory phase by progesterone acting at both the transcriptional and posttranscriptional level (86). The presence of PR is considered evidence of a functional ER-mediated pathway. The administration of a PR antagonist, mifepristone (RU486), in the early secretory phase (LH+2) has been demonstrated to block the progesterone-induced down-regulation of PR (and ER) in nonpregnant human endometrium (87, 88).

B. Endometrial paracrinology
Progesterone is at a maximum concentration in peripheral blood at the mid secretory phase of the cycle when PR in the epithelial cells is waning. PR is absent in leukocytes throughout the menstrual cycle, and therefore the action of progesterone, both prolonged, heralding pregnancy, or falling, in the absence of pregnancy, will be mediated by the USC. The distinct relationship between the USC and the uNK cell has been described above, and the interactions between the USC and the epithelial cell that are necessary to maintain tissue integrity are detailed below. Critically, it is clear that many epithelial functions are controlled by progesterone in a paracrine manner. Not least among these functions is the ability to synthesize and release natural antimicrobial agents that contribute to the normally sterile nature of the uterus (89, 90). Sterility is partly maintained in the uterus by the many innate defenses of the cervix (91), but also by the ability of the endometrium to express a range of natural antimicrobial agents, specific to different phases of the menstrual cycle (89).

A series of experiments by Cunha’s group (92, 93, 94), using knockout mice and tissue recombination, have elegantly shown that both growth and PR expression in epithelial cells is dependent on stromal-epithelial interaction. Estrogen action on the ER receptor in the stromal cells is responsible for the down-regulation of the PR in the epithelial cells (92, 93, 94). Moreover, many of the implantation-permissive changes in murine endometrium have been attributed to optimum estrogen levels (95). Translation of these findings into women needs caution, and evidence from in vitro fertilization programs suggests that estrogen levels in women may not be so critical (95).

C. Endometrial intracrinology
Modification or catabolism of steroids at the level of the target tissue has been described as intracrinology (96). In human endometrium, steroidal regulation of receptor action is dependent on ligand availability. Hence, in reproductive tissues the local actions of sex steroids, estrogens, progestogens, and androgens, is modulated by hydroxysteroid dehydrogenase (HSD) enzymes. The various dehydrogenases are multigene families. The human 17ß HSD family has at least six known members, each being a separate gene product from a different chromosome with distinct properties in terms of substrates and redox direction (96, 97). The type 2 enzyme (17ß HSD-2) plays a major role in inactivation of estradiol to estrone (98). The enzyme 17ß HSD-2 also inactivates testosterone to androstenedione and converts inactive 20-dihydroprogesterone to active progesterone (96, 97). 17ß HSD-2 is expressed in the endometrial glandular epithelium and is up-regulated by progesterone (87). Its activity decreases when progesterone concentrations decrease (as with luteal regression) or after antiprogestogen administration (87, 98).

D. Progesterone and progesterone receptors
Progesterone is essential for the transformation of an estrogen-primed endometrium in preparation for implantation. The molecular and cellular mechanisms by which the sex steroid hormones promote uterine receptivity remain poorly understood. It is, however, recognized that sex steroids, acting via their cognate receptors, initiate a pattern of gene expression essential for implantation and the early stages of pregnancy.

There are two main isoforms of the human PR (71): PRA (Mr 94,000) and PRB (Mr 120,000) arising from a single gene with specific promoters for the two isoforms (99). These function as specific transcriptional regulators of progestin-responsive genes. A third, truncated form (PR-C) (Mr 60,000) can also migrate to the nucleus after steroid activation and may act as a repressor of PRA and PRB (100). PR-C was identified in the breast cancer cell line T47D but has also been reported in the uterus (100). Another molecule (PR-M) with homology to the nuclear receptors mentioned above, lacks a DNA binding domain and therefore may function as a membrane-associated receptor (101).

PRA is the shorter subtype, devoid of 164 amino acids present at the N terminus of the B subtype. It is otherwise identical to the B subtype (102). A significant decline in PR expression in the glands of the functional layer of the endometrium (the upper two thirds of endometrium region that is shed at menstruation) with the transition from the proliferative to the secretory phase of the cycle is well described (9, 84). In contrast, PR expression persists in the stroma in the upper functional region, being particularly highly expressed in stromal cells in close proximity to the uterine vasculature. The basal layer is differentially regulated in that the glands and stroma of the deeper zones express PR throughout the cycle (9). These differences between the superficial and basal layers of the endometrium are likely to be functionally important because only the upper functional zone is shed at menstruation. Localization studies utilizing antibodies that recognize both PR subtypes have described differential regulation of PR in the endometrial epithelium and stromal cells (103, 104, 105). In the secretory phase, the PRB subtype appears to decline in both the stroma and glands, and there is agreement that PRA is the predominant isoform in stroma throughout the cycle (103, 104, 105, 106).

Study of PRA- and PRB-null mice (72) has provided insight into the roles of PR isoforms. In the PRA knockout mouse, estrogen treatment induces uterine epithelial hyperplasia that progesterone treatment cannot suppress. This indicates that the progesterone-mediated suppression of epithelial growth stimulated by estrogen depends on PRA, not PRB. Furthermore, in the PRA+PRB-null mouse, there is a dramatic traffic of inflammatory leukocytes into the uterus, which cannot be prevented by progesterone (107). This implicates a role for progesterone in the suppression of the influx of inflammatory cells into the uterus in wild-type animals. Furthermore, by selective ablation of PRA in mice, it has also been shown that the PRB isoform modulates a subset of reproductive functions of progesterone, by regulation of a subset of progesterone-responsive target genes (108). PRA and PRB are therefore functionally distinct mediators of progesterone action in vivo. It is still not known if these observations in mice can be extrapolated to reproductive function in the human.

A valuable insight about progesterone (and exogenous progestogen) action in human endometrial function may be derived from the observations of pharmacological withdrawal of progesterone from the endometrium (109). Consequently, studies that address the actions of PR antagonists have informed the knowledge base about the local mechanisms that may be targeted to maximize the contragestive and abortifacient properties of these compounds. For example, the antiprogestogen, mifepristone (RU486), is known to exert its inhibitory effects by impairing the gene regulatory activity of the PR (110). Administration of the antiprogestogen, mifepristone, has become a useful model to study local events in both nonpregnant endometrium and early pregnancy decidua in vivo. Examples include studies on the antagonism of progesterone action at the level of its receptor that result in an up-regulation of key local inflammatory mediators (chemokines and prostaglandins) and an influx of leukocytes (24, 111).

Evidence for those functions in the nonpregnant endometrium regulated by progesterone may be derived from in vivo studies where antiprogestogens have been administered acutely in the secretory phase of the cycle or chronically at a low dose. Administration of an antiprogestogen in the early secretory phase will adversely affect local factors of potential importance to implantation, whereas administration in the mid secretory phase will influence factors implicated in endometrial bleeding (109). An increase in steroid receptors (ER, PR, and AR) in both the glandular and stromal compartments in mid secretory phase endometrium after early secretory phase (on day LH+2) administration of antiprogestogens has been described by several authors (85, 87, 88, 112); whether this is due to a failure to down-regulate these receptors has yet to be determined. The endometrial changes (including marked alterations in the endometrial vasculature, Ref.113) associated with withdrawal of progesterone and menstrual bleeding supports the involvement of vasoactive local mediators in this process.

The chronic administration of low-dose oral mifepristone inhibits ovulation and induces amenorrhea or a marked reduction in endometrial bleeding (114), revealing the sensitivity of endometrial morphology to antiprogestogen exposure. Chronic antiprogestogen administration inhibits both endometrial secretion and proliferation and has some intriguing "endometrial antiproliferative effects" (115) that likely involve the AR.

E. Estrogen and estrogen receptors
The precise molecular mechanisms regulated by estrogen in the uterus have not yet been fully defined. Two structurally related subtypes of ER, commonly known as ER and ERß, have been identified in the human, as well as in other mammals (73, 74), and reviewed by Saunders and Critchley (116). The ERß gene, like ER, is encoded by eight exons with maximum levels of homology between ER and ERß present in the DNA and LBDs (75). The function of ERß in the uterus is still not fully elucidated. In both the human and nonhuman primate endometrium, ERß, like ER, is expressed in the nuclei of glandular epithelial and stromal cells and has been reported to decline in the late secretory phase in the functionalis layer (1). However, unlike ER, ERß has been detected with both polyclonal and monoclonal anti-ERß antibodies in the nuclei of the vascular endothelial cells. The presence of ERß in endometrial endothelial cells indicates that estrogen may act directly on endometrial blood vessels (117, 118). Estrogen may therefore have direct effects on endometrial angiogenesis and vascular permeability changes during the cycle. Thus far, PR is reportedly absent from the vascular endothelium (117, 119) of the spiral arteries. The effect of progesterone withdrawal on these vessels, which plays a key role in menstrual induction, is likely to be indirectly mediated by the PR-positive perivascular stromal cells.

In vitro studies have recently demonstrated that homodimers (ER-ER or ERß-ERß) or heterodimers (ER-ERß) may be formed when both isoforms are expressed in the same cell (120, 121). The amount and pattern of expression of each ER subtype is likely to influence gene transcription within that cell. It has been reported that mRNAs encoding isoforms of human ERß formed by alternative splicing of the last (eighth coding) exon are expressed in human tissues (122, 123). Both the mRNA and protein corresponding to one of these splice variants (ERßcx/ß2) are expressed in human endometrium (124). This splice variant lacks the ligand binding site and may act as a negative inhibitor of ERß action (122).

A detailed and thorough inclusive investigation of the nuclear receptor and cofactor mRNA levels in human endometrium and myometrium has recently been conducted by Vienonen et al. (125) and accompanying editorial (126). This study utilizing real-time quantitative PCR has confirmed previous reports describing the menstrual cycle-dependent regulation of expression of endometrial sex steroid receptors. The expression of coactivators was not observed to be regulated.

Retinoic acid receptor (RAR) isoforms may be implicated in the action of progesterone during the secretory phase of the cycle (127). RAR mRNA expression has been reported as more abundant in the proliferative phase of the menstrual cycle, and this follows the cycle-dependent regulation pattern of other sex steroid receptors (ER, ERß, and PR). At the protein level, all isoforms of the RAR, , ß, and , have been described as maximal in the late proliferative phase and to decline in the secretory phase (128). The role of the RAR in human endometrium is yet to be determined.

F. Androgens and the androgen receptor
The AR is expressed in human endometrium (85, 129). This reproductive tissue is a target for androgen action either directly via the AR or indirectly via the ER after aromatization to estrogen (130). Circulating concentrations of testosterone have been reported to show little if any changes throughout the menstrual cycle (in contrast to the cyclical variations in estradiol and progesterone). Testosterone levels are, however, approximately 10 times greater than those of estradiol (131, 132). During the menstrual cycle, the AR is expressed predominantly in the endometrial stroma, and there is considerably higher intensity of AR immunostaining during the proliferative compared with the secretory phase (85). Treatment with androgen will suppress estrogen action in the endometrium, and this effect is most likely mediated by endometrial AR. The physiological role, if any, for AR in the menstrual process is yet to be ascertained. Furthermore, the regulation of AR expression is unknown.

In a clinical situation of chronic hyperandrogenism associated with poor reproductive outcome, polycystic ovarian syndrome (PCOS), there is an elevation of expression of endometrial AR (133). The increase in AR expression was observed in the glandular and luminal epithelium. It is of note that the endometrium from women with PCOS also displays aberrant expression of a proposed biomarker for uterine receptivity, vß3. The expression of this integrin is modified by estrogen and androgens (133, 134). Endometrial epithelial AR is up-regulated by estrogens and androgens in vitro. Expression in vitro is inhibited by progestins and EGF (133).

There is also an intriguing up-regulation of the AR in both glandular and stromal cells after administration of antiprogestogen in both human and nonhuman primate endometrium (85). Androgens do suppress estrogen-dependent endometrial proliferation, and Brenner et al. (115, 135) hypothesized that the endometrial AR is involved with the antiproliferative effects induced by antiprogestogens. Indeed, Brenner’s group (136) has demonstrated in the rhesus macaque that the administration of an antiandrogen, flutamide, will counteract the suppressive effects produced by antiprogestogens on endometrial thickness, stromal compaction, and mitotic index. The endometrial AR may be a critical component of the mechanism by which antiprogestogens suppress endometrial proliferation in the presence of circulating estrogens. Chronic antiprogestogen administration inhibits both endometrial secretion and proliferation (antiestrogen effects). This effect has been described as a "functional noncompetitive antiestrogenic action" of an antiprogestin (137). Because only the endometrial epithelium demonstrates this phenomenon, it has been termed an "endometrial antiproliferative effect" of antiprogestogens (115, 135).

G. Glucocorticoids and the glucocorticoid receptor
Glucocorticoids have been shown to exert specific effects on endometrial cells (138, 139, 140, 141), but their role in endometrial physiology is not defined. Bamberger et al. (142) have briefly described the localization of GR across the menstrual cycle. The GR is almost exclusively expressed in the stromal compartment including endothelial and lymphoid cells (33, 142). Confirmation for an absence of cycle-dependent expression of GR mRNA expression is provided in the recent data on endometrial mRNA levels across the cycle (125). It has been demonstrated both at the mRNA and protein levels that uNK cells express GR (33). The role of glucocorticoids in endometrial immune function remains to be extensively studied. Elsewhere in the body, the immunosuppressive effects of glucocorticoids have led to their wide application in the treatment of inflammatory states. Suggested roles in the uterus for glucocorticoids include effects on implantation (138), endometrial cellular proliferation (139), apoptosis (140), and endometrial remodeling (70). Glucocorticoids have also been shown to repress the decidual prolactin promoter (143) and corticotropin-releasing hormone promoter (144), both of which are markers of decidualization. This and the expression of GR in the endometrium (142) may implicate glucocorticoids in the process of decidualization.

Glucocorticoid function is regulated not only by GR expression but also by the expression of steroid-metabolizing enzymes (which determine the availability of the ligand). The 11ß HSD family modulates the action of glucocorticoids by converting either cortisone (inactive) to cortisol (11ß HSD-1) or cortisol (active) to cortisone (11ß HSD-2). Smith (62) reported that levels of the 11ß HSD-2 are higher across the nonpregnant menstrual cycle than 11ß HSD-1 and that 11ß HSD-2 was present in the luminal and glandular epithelium with raised levels in the secretory phase. Hence, it was suggested that the balance of expression of 11ß HSD isoforms could facilitate trophoblast invasion by removing the glucocorticoid-mediated inhibition of MMPs. In this context, data have been reported on the expression of 11ß-HSDs during in vitro decidualization of human endometrial stromal cells (145). Decidualization was observed to involve an enhancement of the corticosteroid-metabolizing capacity of stromal cells, thereby implicating a mechanism by which stromal cells might influence the health and invasiveness of the implanting trophoblast.

Detailed in vivo studies of 11ß HSD-1 mRNA and protein levels in very early pregnancy tissues have not to our knowledge been reported. To date, there is no commercially available antibody for the immunolocalization of the 11ß HSD-1 enzyme. It is therefore of particular interest that uNK cells that express GR mRNA and protein are found aggregated close to the glandular epithelium and also have proposed roles in the control of early trophoblast invasion (146).

	IV. Progesterone Withdrawal and the Mechanisms of Menstrual Bleeding

Top Abstract I. Introduction II. Endometrial Morphology III. Steroid Control in... IV. Progesterone Withdrawal and... V. Repair and Vessel... VI. Disorders of Menstruation VII. Local Mediators Associated... VIII. Perspective and Future... References

 
A. Progesterone action in the endometrium
Several array-based studies have examined the changes in gene expression associated with the mid secretory phase of the menstrual cycle (5, 6, 147), and these have recently been reviewed by Dey et al. (148). The mid secretory phase of the cycle is a critical time because it represents the height of progesterone priming, providing both the implantation window and heightened sensitivity to progesterone withdrawal. Thus, changes induced at this time will be critical to effective implantation of the blastocyst and, in the absence of pregnancy, to menstruation. These studies have shown good agreement in reporting important gene changes around this time. These investigations of the implantation window have been supported by studies on the induction of the decidualized phenotype in endometrial stromal cells (149, 150). Because this system is more readily manipulable, time courses of gene expression have been possible (149), and these show very rapid increases in well-identified markers of decidualization such as IGF binding protein (IGFBP)-1.

Some changes seen in the mid secretory phase of the cycle such as the expression of Dickkopf-1 (Dkk-I) (5, 147, 151) are novel and revealing. Dkk-I has been established as an inhibitor of the WNT signaling pathway, which is active in human endometrium (152). Although most WNT gene expression does not change through the cycle, the inhibitor Dkk-I in stromal tissue is progesterone dependent (152). WNT-3 was the one gene modulated, with elevated expression during the proliferative phase of the cycle, and it is WNT-3 that is implicated in increased cyclooxygenase (COX)-2 expression and prostaglandin E₂ (PGE₂) synthesis in mammary epithelial cells (153).

Dkk-I inhibits a coreceptor for the Frizzled receptor low-density lipoprotein associated receptor-6, and thus a potential new pathway of progesterone control of endometrial differentiation is apparent (152). Frizzled in turn is implicated in HOX gene control. HOX genes are best characterized as developmental signals, with HOX-10 playing an important role in uterine differentiation (154), but HOX-10 in particular is essential for implantation (155, 156), and pathologies such as endometriosis, polycystic ovarian disease, and the presence of an hydrosalpinx are all associated with aberrant HOX-10 expression (157, 158, 159). One role for HOX-10 has been identified as the mediator of progesterone-controlled expression of the prostaglandin receptors E-series prostanoid (EP)3 and EP4 (160).

Additional examples of endometrial genes that are regulated by progesterone include glandular secretion of glycodelin (otherwise known as pregnancy protein 14, progestagen-dependent endometrial protein, or 2 pregnancy-associated endometrial globulin) (161), 15-hydroxyprostaglandin dehydrogenase (PGDH) (162, 163, 164), 17ß HSD-2 (87), prolactin (46), and calcitonin. The secretion of glycodelin is of interest because endometrial and blood levels are maximal 10 to 15 d after the LH surge (165). Thus, levels remain high (or even increase) after the levels of progesterone begin to fall; explanations for this are made more difficult because of the pleiotropic nature of glycodelin. However, the most likely role for glycodelin in endometrium is as an epithelial morphogen (166, 167), and as such the effect of progesterone is likely to be indirect through the influence of stromal cells. It is thus possible that further differentiation of glandular epithelium is a prerequisite to menstruation.

Expression of calcitonin mRNA has been demonstrated to be temporally restricted to the mid secretory phase of the cycle, a period that coincides with the putative window of implantation (168). The site of postovulatory synthesis of calcitonin mRNA and protein is the glandular epithelium. Evidence for regulation of this gene by progesterone has been derived from examination of endometrium collected from women treated with an antiprogestogen, mifepristone. Calcitonin expression was dramatically reduced in women exposed to acute administration of mifepristone in the early secretory phase (administered day LH+2).

The cellular interactions and progesterone target genes involved in the decidualization process are complex. Multiple growth factors, cytokines, and protein hormones have been recognized as important signals for initiation and maintenance of decidualization (reviewed in Refs.46 and 169). Gene array techniques have helped our understanding of the uterine changes in the implantation window and their progesterone dependence. Most importantly, the decidualized stromal cell, by virtue of its retained PR, is likely to be a cell that is critically affected by falling progesterone, signaling the onset of menstruation.

The disturbed bleeding patterns reported by women with the use of progestogen-only contraceptives are likely to reflect modifications in the endometrial vessels from which bleeding arises, including changes in vessel integrity and/or hemostasis. Furthermore, the aberrant bleeding may arise from a different vascular source than normal menstrual bleeding (170). Breakthrough bleeding arises mainly from capillaries and veins adjacent to the uterine lumen and considered to be related to increased vessel fragility (171, 172, 173).

The levonorgestrel (LNG)-releasing intrauterine system (LNG-IUS) is now widely used for the management of heavy uterine bleeding, although its principal indication for use is contraception (174). The intrauterine delivery of LNG induces a rapid and dramatic transformation of the endometrium, characterized by extensive decidualization (175, 176). The observed morphological changes are consistent with progesterone-mediated differentiation as observed during the progesterone-dominated secretory phase and during pregnancy (32). With local intrauterine LNG administration, there is no longer cyclical activity within the endometrium and there is a general thinning of the functional layer of the endometrium. The features of atrophy and decidualization are evident within 1 month of insertion of the LNG-IUS. Importantly, the morphology of the endometrium returns to normal within 1–3 months of the removal of the device, and there is a complete return of previous fertility (177). Initially, after LNG-IUS insertion, sex steroid receptor content is decreased with consequent altered expression of local mediators that may play a role in aberrant bleeding episodes. It is disappointing that no single factor has as yet been identified to explain the mechanism(s) of abnormal bleeding patterns associated with the use of this or indeed any other progestogen-only contraceptive. The endometrial responses to local LNG exposure have been documented and in summary include (reviewed in Ref.178): down-regulation of ER, PR, and AR; expression of prolactin (stroma) and prolactin receptors (epithelium and isolated leukocytes in stroma) and IGFBP-1; elevation of leukocyte infiltrate after insertion (uNK cells, macrophages); enhanced expression of local inflammatory mediators (cytokines and prostaglandins); evidence for aberrant angiogenesis; changes in vessel integrity and/or hemostasis; and abnormally fragile superficial endometrial vessels.

The sc delivery of LNG (Norplant) also has a marked effect on the endometrial vasculature. A decreased expression of a number of components of the endothelial cell basement membrane is evident with Norplant administration (172). These architectural changes are highly likely to play a role in endometrial vessel integrity and fragility. The processes that lead to increased vessel fragility and changes in vessel density are, however, yet to be determined. Indeed, modulation of the vascular basement membrane is likely to be part of a cascade of events that results in aberrant angiogenesis.

B. Decidual changes preceding menstruation
The primary purpose of decidualization is to prepare the endometrium for the implanting blastocyst, but at the same time preparation has to be made for a failed implantation. The depth of the implantation in humans is a characteristic that may necessitate sloughing off a large proportion of the endometrial surface. The changes seen in USCs during the secretory phase will be relevant to the onset of menstruation, particularly as these cells retain the PR. Decidualization of stromal cells can be viewed as differentiation with a concomitant reduction in growth factors induced by an early appearance of agents such as IGFBP-1 and -3 and spermidine/spermine N-acetyl transferase (149, 150, 179). Although IGFBP-1 may have other, non-IGF-related functions (180), it is likely that this binding protein restricts IGF levels in the immediate environment of the secreting cell. IGF is an estrogen-induced growth factor prominent in the proliferative phase (180), and polyamines are also growth factors for endometrial stromal cells (181) whose action is restricted through acetylation by acetyl transferase.

Decidualization is also accompanied by an increase in the secretion of matrix components, particularly such as collagen, fibronectin, and laminin (169). Moreover, agents that degrade matrix such as MMP-3 have to be maintained at low levels by a progesterone and IL-1-mediated mechanism to allow decidualization (182). Cytokine changes that suggest paracrine actions of the decidual cell also occur during decidualization; the decidual cell secretes IL-15, which is an essential growth and differentiation factor for the uNK cell (40, 41, 183) (also see Section II.C).

C. Physiological withdrawal of progesterone
The withdrawal of progesterone prevents implantation and converts the refractory pregnant uterus once again into a spontaneously steroid responsive organ (184, 185). The physiological withdrawal of progesterone from an estrogen-progesterone primed endometrium (that occurs with demise of the corpus luteum due to the absence of pregnancy) is also the triggering event for the cascade of molecular and cellular interactions that result in menstrual bleeding. A current hypothesis for menstruation (described below) is based on lines of evidence derived from studies on local endometrial response to progesterone withdrawal (1).

The withdrawal of progesterone up-regulates key inflammatory mediators, many of which have a key perivascular location (25, 186, 187), underlining the role of the stromal cell. Among the agents stimulated are chemokines: the -chemokine CXCL8 (neutrophil chemotactic factor, IL-8) and the ß-chemokine CCL-2 (monocyte chemotactic peptide-1, MCP-1), as well as the inducible enzyme responsible for synthesis of prostaglandins, COX-2 (111, 188).

Early studies of menstruation implicated prostaglandins (189) and accord with both increases in prostaglandin synthesis and decreases in metabolism in response to falling progesterone levels (190). Prostaglandin synthesis via COX-2 is particularly relevant in the vascular compartment because this provides an explanation for the action of nonsteroidal antiinflammatory drugs (which inhibit COX enzyme activity) in menstrual pathology. Moreover, the actions of prostaglandins on blood vessels and surrounding cells is underlined by the significant distribution of prostaglandin receptors in this locus (Fig. 3) (191, 192). PGDH, the enzyme responsible for conversion of prostaglandins to inactive metabolites, is a progesterone-dependent enzyme (163, 164). Antagonism of progesterone action results in an inhibition of PGDH expression, and progesterone withdrawal from decidua clearly demonstrates a perivascular locus for this enzyme (193).

View larger version (75K): [in this window] [in a new window]   FIG. 3. Localization of EP2 (A), EP4 (B), and FP (C) receptors in the vascular compartment (arrowhead) of human endometrium.

2

It has long been accepted that myometrial contractions and vasoconstriction are a consequence of an increased production of PGF₂, consequent upon progesterone withdrawal (190). Coincident vascoconstriction of the endometrial spiral arteries takes place (2), and so the uppermost endometrial zones are presumed to become hypoxic. Hypoxia is a potent inducer of angiogenic and vascular permeability factors such as VEGF (202), and a hypoxia-dependent mechanism to initiate menstruation is attractive, but there is however controversy concerning the role for hypoxia (if any) in the menstrual process (203).

The angiogenic factor, VEGF is a local mediator stimulated by hypoxia in endometrial stromal cells (63). Progesterone withdrawal has been reported to up-regulate the endometrial stromal expression of the VEGF type 2 receptor, kinase domain receptor (KDR), in women and nonhuman primates (204). This stromal but not vascular endothelial expression of KDR is blocked by adding back progesterone 24 h after progesterone withdrawal. Pro-MMP-1 is also up-regulated in a coordinate manner in the same stromal cell population by withdrawal of progesterone. Furthermore, Nayak and Brenner (205) have described the up-regulation of VEGF-A mRNA in the glands and stroma of the same superficial endometrial zones. Hence, given that VEGF-A, KDR, and MMPs are coordinately expressed by stromal cells of the upper zones of premenstrual stage endometrium at the time of progesterone withdrawal, the conclusion is that a VEGF-KDR-MMP link is an important component of the premenstrual/menstrual process (115, 117, 204).

D. Menstruation, the sequence
Thus, early events occurring in PR-positive cells herald the onset of menstruation but may be inhibited by "add back" of progesterone. In the rhesus macaque monkey, the adding back of progesterone before 36 h following progesterone withdrawal prevented menstrual bleeding (206). However, add back of progesterone after 36 h was ineffective in preventing the onset of menstruation. Thus, the withdrawal of progesterone will initially affect cells expressing the PR in a reversible manner. These early, progesterone withdrawal events in menstruation involve vasoconstriction and cytokine changes (207). Subsequent events are likely to be irreversible and include the activation of lytic mechanisms in a cascade of activation of pro-MMPs and accentuation by hypoxia. Hence, the latter phase of menstruation is progesterone independent and will involve cells that may not express the PR. These changes will involve the disruption of the progesterone-dominated epithelial-stromal interaction that suppresses key mediators such as IL-1, MMP-1 (208), and MMP-7 (209). IL-1 may be particularly important because this cytokine has far-reaching effects needing tight control by the multiple pathways that include posttranslational modifications, decoy receptors, and the receptor antagonist (IL-1 receptor antagonist) (Fig. 4).

View larger version (30K): [in this window] [in a new window]   FIG. 4. Progesterone withdrawal activates many pathways; prominent among these are those releasing vasoactive agents. Chemotactic agents, which synergize with vasoactive agents, are also expressed. There is no doubt that MMPs, which degrade the interstitial matrix of the endometrium, play a major role in the menstruation process, but their origin is uncertain. MMPs can be released from both resident endometrium stromal cells and invading leukocytes.

The roles of the leukocyte and the resident stromal cell scenarios are not mutually exclusive because early progesterone-related events are dependent on the PR, which is absent from leukocytes. Moreover, local release of agents such as MMP-1 and MMP-3 from the stromal cell (210) may activate other proteases released from the invading neutrophils. This is supported by the presence of substantially more latent MMP-9 than active MMP-9 before menstruation (211).

1. MMPs from stromal cells.
In general, MMPs in uterine cells are repressed by progesterone, but this is not always direct and may involve interaction with the stromal cell. Indirect action is likely to affect MMPs from both epithelial and endothelial cells. In epithelial cells, MMP-7 is suppressed at a time of high progesterone concentrations because of the release of TGF-ß from stromal cells in response to progesterone (209). The endothelial cell may also respond to TGFß-1 signaling because it does not possess the nuclear PR at any stage in the cycle. These interactions underline the critical role of the stromal cell in the initiation of menstruation. By the late secretory phase, the spiral arterioles are surrounded by a characteristic cuff of distinct cells expressing smooth muscle characteristics (e.g., smooth muscle actin expression) in common with decidual cells (23). Withdrawal of progesterone from decidua shows that these distinctive cells release proinflammatory agents in a manner that is clearly different from surrounding decidual cells (24, 25).

Observation of changes in vivo suggests progesterone-dependent suppression of collagenase and other lytic enzymes in the uterus (212) and in particular the suppression of MMP-1 and MMP-3 (213) and MMP-9 (214). The effects of progesterone on the lytic enzymes may be enhanced by progesterone-stimulating synthesis of tissue inhibitors of MMPs (TIMPs), such as TIMP-3 (215). Although mRNA for MMP-1 and MMP-3 can be detected in proliferative phase endometrium by in situ analysis, this disappears during the secretory phase. Moreover, when progesterone levels decline in the late secretory phase, expression of MMP-1 and MMP-3 is reinitiated. These findings are supported by evidence from culture studies showing that MMP-1 expression and protein release were inhibited in the presence of progesterone (216).

Initial steps in the degradation of collagen are ascribed to MMP-1, and thus this enzyme is critical to the stability of the basement membrane of the blood vessels. After initial cleavage by MMP-1, other MMPs will contribute including MMP-2, which, although expressed throughout the menstrual cycle in stromal cells (217), rises in response to the withdrawal of progesterone (218).

Because MMP-1 plays a key role in extracellular matrix breakdown, tight control of the protein is necessary. This is achieved by transcriptional control as well as control of the stability of the mRNA. The coding region for MMP-1 contains both activator protein-1 and nuclear factor B (NFB) response elements (219) that allow stimulation not only by inflammatory agents such as IL-1 and TNF but also by progesterone.

MMP-1 and MMP-3 have similar promoter regions and in this respect differ from MMP-2 (220). NFB is important in control of MMP-1, MMP-3, and MMP-9 and, when activated along with other transcription factors, can stimulate many other proinflammatory genes such as inducible nitric oxide synthase, TNF, IL-1ß, toll-like receptor-4, and COX-2 (221, 222). NFB may be controlled by progesterone in several ways: progesterone may stimulate inhibitor of B, the protein that retains NFB outside the nucleus (223), or the PR may compete with binding sites for NFB on promoter regions of the gene (224). However, the real key to the importance of NFB is the finding that TGFß-1 and progesterone may have a coordinate suppressive effect (225) that may be through NFB because of the interactions of both with this pathway.

Suppression of MMP-1 production is affected by TGFß-1 through a SMAD3/4-dependent mechanism (226) acting on NFB. This inhibition occurs by competition between SMAD3 and the NFB complex for P300, which is a transcriptional coactivator for both (226). This results in an inhibition of the acetylation of key lysine residues that would normally render the NFB complex immune to inhibition by inhibitor of B (226).

In addition, apart from suppressing MMP activity, TGFß-1 is also responsible for the stimulation of expression of TIMPs (227) and increasing the synthesis of major matrix proteins such as collagen and fibronectin (228). The other function of TGFß is the control of cell growth, and epithelial cell growth is inhibited by TGFß (229). Thus, TGFß-1 stabilizes tissue by limiting MMP activity, which accords with the cyclical expression of TGFß with the highest levels in the secretory phase stromal cell (230, 231). Moreover, TGFß-1 has to be activated in a proteolytic step, although little is known about the physiological agents involved. One candidate lytic enzyme is uPA, which is reduced in the mid secretory phase (231) but is expressed when progesterone levels fall before menstruation. uPA activity in turn is inhibited by tissue factor (TF), which is progesterone dependent (232) and appears in stromal cells in the secretory phase of the cycle and in decidua (233). Expression of TF, which in decidualized endometrial stromal cells is both delayed and chronic (234), is largely controlled by the specificity protein (SP)1 transcription factor (235). In vitro studies show that progesterone stimulates SP1 and inhibits SP3, which antagonizes SP1, and that SP1 is enhanced and SP3 ablated in perivascular cells in the secretory phase of the menstrual cycle (235). Because TF is a major hemostatic agent (236), it is beneficial in perivascular cells in that potential bleeding will be limited around the time of implantation and later when extravillous trophoblast cells invade the maternal arteries. However, TF levels fall before menstruation to allow menstrual-associated hemorrhage (237).

Another relevant gene with an SP1 response element is plasminogen activator inhibitor-1 (PAI-1) which inhibits the fibrinolytic pathway. Thus, progesterone stimulates PAI-1 expression in endometrial stromal cells, possibly moderating decidual cell migration within tissue. Trophoblast certainly expresses plasminogen activator that is inhibited by a PAI-1-vitronectin complex, and thus the expression of PAI-1 by decidua may be a mechanism for the restriction of trophoblast invasion. With the decline of progesterone, PAI-1 expression will be restricted, and thus fibrinolytic activity will be present at the time of menstruation, which will account for the reduced clotting in menstrual blood.

2. MMPs from leukocytes.
Invasion of the endometrium by leukocytes is an integral part of the process of menstruation, both contributing to tissue breakdown and repair (see Section I.C). Leukocytes that are attracted into the uterus before menstruation are a major source of lytic enzymes. The neutrophil in particular represents an almost unlimited source of MMP-8 (neutrophil collagenase) and MMP-9 (gelatinase B). The neutrophil characteristically has a high turnover rate and a high production rate of 10¹¹ cells per day with any deficiency quickly replenished by the bone marrow (