Key words: endometriosis; menstruation; preconditioning; preeclampsia; pregnancy

Article Outline

The link between menstruation and placentation

Uterine vascular remodeling in normal pregnancy and preeclampsia

Prepregnancy preparation for pregnancy

Menstrual preconditioning, endometriosis, and preeclampsia

Clinical Perspective

Further Reading

Appendix

References

Eutherian (placental) mammals have evolved different strategies to ensure reproductive success. The type of placenta and the depth of trophoblast invasion among various primate species appear to correlate with the extent of endometrial remodeling prior to conception and the occurrence of menstruation.¹ For instance, strepsirrhine primates (lemurs and lorisiforms) have noninvasive epitheliochorial placentae and completely lack menstruation whereas simian primates (Old World monkeys, apes, and human beings) uniformly posses invasive hemochorial placentae and typically exhibit some form of menstruation, although there is considerable variation in the degree of blood loss.¹ Compared with our primate relatives, menstruation in human beings is not only extraordinarily heavy but placentation is also exceptionally deep, with trophoblast invading not only the decidual endometrium but also the inner third of the myometrium, termed the “uterine junctional zone,” and its spiral arteries.^{[2] and [3]}

Endometriosis is foremost a menstrual disorder, characterized by the presence of hormone-dependent ectopic endometrial implants, impaired differentiation of the eutopic endometrium, and thickening of the uterine junctional zone.^{[4] and [5]} As endometriosis and preeclampsia are disorders at opposite ends of the reproductive process, there may be few reasons to suspect that both disease processes are interrelated. Yet, a recent case-control study reported that women with endometriosis-related infertility receiving in vitro fertilization treatment have a markedly reduced risk of developing preeclampsia or pregnancy-induced hypertension.⁶ Admittedly, the protective effect of endometriosis against preeclampsia or other obstetric disorders associated with impaired deep placentation–such as fetal growth restriction, placental abruption, and preterm labor–requires further validation in prospective studies. Nevertheless, these observations do raise questions regarding the role of cyclic endometrial remodeling and menstruation in determining subsequent pregnancy outcome.

The link between menstruation and placentation

Menstruation is defined by shedding of the superficial endometrial layer in response to decreasing progesterone levels accompanied by overt bleeding. Notably, uterine bleeding associated with ovulation or intercourse occurs in a variety of species but does not constitute menstrual bleeding. Apart from simian primates, true menstruation is found only in certain bat species, such as wild fulvous fruit bats, and perhaps the elephant shrew.^{[7] and [8]} One unifying feature of the endometrium of menstruating species, including bats, appears to be the profound progesterone-driven remodeling of the stromal compartment, a process termed “decidualization” (Figure, A and B). This differentiation process is primarily defined by the transformation of stromal fibroblast into secretory epithelioid-like decidual cells and is further characterized by massive influx of specialized immune cells, especially uterine natural killer (uNK) cells, and vascular remodeling. Decidualization, however, also represents a tipping point, after which the integrity of the endometrium becomes inextricably dependent on elevated progesterone levels (Figure, C).^{[7] and [9]} Thus, although all mammals with invasive placentae exhibit an endometrial decidual response on embryo implantation, only in menstruating species is this process directly under maternal control and occurs spontaneously during the late secretory phase of each cycle.

Menstruation is generally viewed as a nonadaptive consequence of uterine evolution, a concept first proposed by Finn.¹⁰ In other words, menstruation may have emerged as an inevitable consequence of the process of extensive endometrial remodeling necessary to accommodate deep hemochorial placentation but serves no other purpose except to reinitiate the endometrial cycle to allow a future pregnancy. Yet, it is striking that human reproduction has evolved in ways that limit the likelihood of pregnancy while maximizing the frequency of menstruation. For example, in comparison with many mammals with rapid reproductive turnover, ovulation in human beings and other primates occurs spontaneously rather than triggered by mating. Human beings also do not exhibit overt estrous behavior, have a limited and variable fertile window and concealed ovulation, lack embryonic diapause, and have a high incidence of preimplantation embryo wastage. Moreover, 90% of all menstrual cycles in very young adolescent girls are anovulatory.¹¹

At first glance, the convergence of all these reproductive features in a single species, human beings, seems to be a pretty unfortunate evolutionary coincidence, accounting on the one hand for disappointingly low monthly fecundity rates ( 20%) and on the other for the high frequency of undesirable menstrual events. However, there are reasons to challenge this view of human reproduction as an evolutionary accident. In fact, the reciprocal relationship between fecundity and menstruation suggests that repeated menstrual priming or preconditioning of uterine tissues and its vasculature during nonconception cycles could be instrumental for successful deep placentation. This conjecture fits well with the observation that preeclampsia, a disorder characterized primarily by absence of spiral artery remodeling in the junctional zone myometrium,^{[3] and [12]} is more prevalent and severe in a first pregnancy and especially so in young women. For example, data from the National Hospital Discharge Survey in the United States identified a maternal age < 20 years as the strongest risk factor for both preeclampsia and eclampsia.¹³ Conversely, it is well established that the risk of preeclampsia decreases sharply after the first full-term pregnancy and, albeit somewhat controversially, after a miscarriage.

Uterine vascular remodeling in normal pregnancy and preeclampsia

To answer the question of whether events in the menstrual cycle impact on pregnancy outcome, it is useful to first briefly revisit the uterine vascular changes associated with deep hemochorial placentation. Blood supply to a 50-g nonpregnant human uterus represents < 1% of the cardiac output. By the end of gestation, uterine perfusion has increased to 25% of cardiac output, 90% of which is carried through a limited number of physiologically transformed spiral arterioles into the intervillous space of the placenta. Thus, from a maternal perspective, pregnancy is an example of extraordinary rapid histogenesis, which is unrivalled in normal adult tissues.

Whether or not a placenta meets the fetal requirements for respiratory gas exchange and nutrient supply depends entirely on coordinated interstitial and endovascular invasion of the maternal decidua and junctional zone myometrium by extravillous cytotrophoblast (Figure, D and E). Endovascular invasion is preceded by interstitial trophoblast migration and is indispensable for the destruction of the arteriolar structure of the spiral arteries and the acquisition of a distended fibrofibrinoid vessel structure in these uterine zones.^{[14] and [15]} Initially, this process is characterized by plugging of the terminal decidual vessels by endovascular trophoblast, thereby restricting the exposure of the early placenta and fetus to maternal arterial blood. Subsequently, the trophoblast plugs are dislocated, allowing perfusion of the placental intervillous space. This transition occurs toward the end of the first trimester and is thought to be associated with profound changes in the oxygen tension at the fetomaternal interface.¹⁶ The endovascular trophoblast continues to invade the spiral arteries, replaces the endothelial lining temporarily, and eliminates most of the musculoelastic tissue in the vessel. The morphological features of this process of physiological transformation of the spiral arteries have been extensively described elsewhere.^{[2], [14] and [17]} Classically, 2 phases have been described: a deciduotrophoblast phase completed by 10 weeks gestation that involves remodeling of the endometrial portion of the spiral arteries and a myometriotrophoblast phase, from 14 weeks onward, encompassing the junctional zone portion of the vessels that critically reduces uterine blood flow resistance necessary for a reciprocal increase in placental perfusion.

Preeclampsia is increasingly attributed to shallow invasion of maternal tissues by extravillous cytotrophoblast.¹⁸ The term “shallow” trophoblast invasion is unfortunate for 2 reasons. First, it seems to imply that preeclampsia is primarily and/or commonly caused by defective cytotrophoblast differentiation and/or invasiveness, for which there is little or no direct evidence. Second, there is also no evidence of shallow invasion of the decidua or the decidual spiral arteries in preeclampsia. However, endovascular trophoblast invasion of the junctional zone portion of the spiral arteries is impaired, causing muscular and sometimes atherotic vessels to remain in this uterine zone (Figure, F).^{[3], [12] and [15]} Although it has been reported that interstitial invasion of the junctional zone myometrium is more superficial in preeclampsia,¹⁹ this is, at least in our experience, far from an universal feature. The reason for impaired deep invasion of the junctional zone spiral arteries is unknown but may result from either an extreme maternal hyperinflammatory response to pregnancy,²⁰ inadequate decidualization,^{[3] and [21]} aberrant uNK cell function or impaired activation on trophoblast interaction,²² increased trophoblast cell death,¹⁹ or a combination of the above.²³

Beyond doubt, however, is the fact that preeclampsia is a disease of early pregnancy. Uterine artery Doppler studies in the first trimester of pregnancy can reportedly identify at least 50% of patients who will subsequently develop preeclampsia.²⁴ Combining Doppler studies with serum markers, such as placental protein 13, may improve prediction, especially of early-onset preeclampsia.^{[25] and [26]} There is also considerable interest in the predictive value of circulating antiangiogenic factors, for example soluble fms-like tyrosine kinase and endoglin, although this approach may require serial measurements during the first and second trimesters of pregnancy.²⁷

Prepregnancy preparation for pregnancy

Although endovascular trophoblast is indispensable for physiological transformation of the spiral arteries, it is important to note that maternal preparations start well before, and independently of, trophoblast invasion. During the late secretory phase of the cycle, decidualizing endometrial stromal cells form a cuff around the changing spiral arteries, characterized by endothelial swelling, vacuolation, and disorganization of the smooth muscle media.²⁸ Biochemically, decidual cells play a decisive role in ensuring tissue hemostasis by expressing the fibrinolysis inhibitor plasminogen activator type 1 and tissue factor, a membrane-anchored glycoprotein that serves as the receptor for coagulation factor VII/VIIa.²⁹ On secretory transformation, decidualizing endometrial cells also acquire the means to respond to trophoblast signals and to provide histotrophic support to the early conceptus.^{[30] and [31]} Local production of chemokines such as CCL4, CXCL9, and CXCL10 triggers influx of specialized (CD56bright/CD16^–) uNK cells, which are a rich source of growth and angiogenic factors essential for vascular remodeling.^{[32] and [33]} These cells are also capable of modulating T-cell function through expression of glycodelin A and galectin-1.³⁴ In pregnancy, immunotolerance toward the conceptus is further achieved by increased activity of the indoleamine 2,3-dioxygenase, a tryptophan-catabolizing enzyme expressed in decidual cells and syncytiotrophoblast.³⁵ Moreover, decidual cells are highly adapted to resist oxidative insults through a number of defense and repair mechanisms.^{[36] and [37]} For instance, a variety of intracellular and extracellular free radical scavengers, such as superoxide dismutase 2, peroxiredoxin 2, thioredoxin, glutaredoxin, and glutathione peroxidase 3, are induced on decidualization. Differentiating stromal cells also express high levels of GADD45α, a factor involved in DNA repair and stress responses. Lastly, the proapoptotic forkhead transcription factor FOXO3a is no longer induced in decidual cells on oxidative stress, thereby preventing activation of the apoptosis machinery responsible for oxidative cell death in undifferentiated stromal cells.³⁷

Taken together, uterine tissues acquire some unique features on decidualization that allow simultaneous protection of the conceptus against environmental or immunological insults and ensure hemostasis and tissue integrity during deep trophoblast invasion. Expression of a decidual phenotype is, however, strictly dependent on elevated progesterone levels. In the absence of pregnancy, decreasing progesterone levels not only reverse the decidual phenotype but also induce the expression of a gene network that encode for chemokines, proinflammatory cytokines, matrix metalloproteinases, and apoptotic factors, leading to influx of inflammatory cells, proteolytic breakdown of the extracellular matrix, cell death, focal bleeding, and menstruation.^{[7] and [38]}

Menstrual preconditioning, endometriosis, and preeclampsia

Menstruation and pregnancy are both inflammatory events, albeit of very different magnitudes. In addition, both events are associated with variable degrees of uterine free radical production, oxidative stress, ischemia-reperfusion injury, vascular remodeling, and angiogenesis. Therefore, it seems not only plausible but likely that cyclic endometrial remodeling and menstruation precondition the human uterus for pregnancy. The term “preconditioning” refers to the paradoxical yet ubiquitous biological phenomenon that a brief exposure to a harmful stimulus at a dose below the threshold for tissue injury provides robust protection against, or tolerance to, the injurious effects of a subsequent more severe insult.39 R.B. Jennings, C.E. Murry, C. Steenbergen Jr and K.A. Reimer, Development of cell injury in sustained acute ischemia, Circulation 82 (1990), pp. II2–II12.^{[39] and [40]} In recent years, the mechanisms underpinning preconditioning have been widely studied because of its therapeutic potential in preventing cardiac or cerebral injury in different clinical settings. Notably, reactive oxygen species production, activation of redox-sensitive signaling pathways, expression of angiogenic factors such as vascular endothelial cell growth factor, and resistance to cell death are major effectors in the process of preconditioning,⁴⁰ all of which are highly regulated in the endometrium in response to hormonal signaling and perturbed in the presence of endometriosis.^{[41], [42], [43] and [44]} Thus, it is tempting to speculate that endometriosis is primarily a disease of exaggerated endometrial preconditioning, which not only confers protection against hyperinflammation and oxidative stress associated with pregnancy but also endows endometrial cells with the mechanisms to survive in unfavorable ectopic locations.

To confer protection during pregnancy, cyclic uterine preconditioning must have prolonged and sustained effects that impact on the cellular responses in the basal endometrial layer, which contain the progenitor cells for the underlying junctional zone and superficial endometrium.⁴⁵ Although direct proof of cyclic menstrual preconditioning is as yet lacking, it is striking that the junctional zone myometrium is significantly thicker on T₂-weighted magnetic resonance imaging in patients with endometriosis when compared with age-matched control subjects.⁵ Power Doppler ultrasound studies have also shown that endometriosis is associated with increased endometrial-subendometrial blood flow during the late secretory phase of the cycle.⁴⁶ In fact, the reciprocal correlation between the degree of uterine perfusion prior to pregnancy and the likelihood of obstetric complications may extend to other reproductive disorders, such as polycystic ovary syndrome and unexplained infertility (Appendix 1).

Perhaps the most compelling evidence for cyclic endometrial preconditioning comes from the observation that stromal cells purified from eutopic endometrial biopsy specimens from patients with and without endometriosis exhibit different responses to a decidualizing stimulus, even after prolonged culture.^{[47] and [48]} Such sustained reprogramming of cellular responses likely involves epigenetic changes like DNA methylation or posttranslational histone tail modifications. This is indeed the case as illustrated by experiments in the baboon model for endometriosis. Kim et al⁴⁹ demonstrated that induction of endometriosis and chronic inflammation in this model resulted in a gradual decrease in endometrial HOXA10 expression, a homeobox transcription factor involved in endometrial development. Importantly, this down-regulation was only significant 6-12 months after the induction of endometriosis, corresponded to increased methylation of the proximal promoter of the HOXA10 gene, and was associated with enhanced expression of decidual markers, such as insulin-like growth factor binding protein-1.⁴⁹

Clinical Perspective

Human reproduction has evolved in a way that inextricably links cyclic menstruation to optimal pregnancy outcome. We propose that cyclic endometrial decidualization and menstrual shedding is an example of physiological preconditioning that prepares uterine tissue for the dramatic vascular remodeling, reactive oxygen species production, and hyperinflammation associated with deep hemochorial placentation. In analogy with other organ systems, modest cyclic inflammation and oxidative stress may have cumulative and sufficiently large effects on the uterine epigenome to trigger an adaptive protective response and to induce changes in uterine hormonal responses and perfusion conducive for pregnancy. Understanding if and how cyclic endometrial remodeling is coupled to pregnancy outcome is more than merely an academic exercise. First, it provides a new paradigm for understanding prevalent reproductive disorders, such as endometriosis, and their impact on pregnancy outcome. Second, it raises the real possibility of identifying those patients at risk of adverse pregnancy outcome prior to conception, on basis of uterine perfusion studies, epigenetic analysis, or endometrial gene expression profiling during the late secretory phase of the cycle. If correct, our preconditioning hypothesis may also lead to novel and perhaps simple strategies to prevent adverse pregnancy outcome in at-risk patients.

Appendix

Drs J.J. Brosens and Parker are supported by the Institute of Obstetrics and Gynecology Trust.

Cite this article as: Brosens JJ, Parker MG, McIndoe A, et al. A role for menstruation in preconditioning the uterus for successful pregnancy. Am J Obstet Gynecol 2009;200:615.e1-615.e6.