This Article Abstract Full Text (PDF) All Versions of this Article: 28/1/117    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 Stocco, C. Articles by Gibori, G. Search for Related Content PubMed PubMed Citation Articles by Stocco, C. Articles by Gibori, G.

The Molecular Control of Corpus Luteum Formation, Function, and Regression

Department of Obstetrics, Gynecology and Reproductive Science (C.S.), Yale University School of Medicine, New Haven, Connecticut 06510; Division of Basic Biomedical Sciences (C.T.), Sanford School of Medicine of the University of South Dakota, Vermillion, South Dakota 57069; and Department of Physiology and Biophysics (G.G.), University of Illinois College of Medicine, Chicago, Illinois 60612

Correspondence: Address all correspondence and requests for reprints to: Dr. Geula Gibori, Department of Physiology and Biophysics, University of Illinois College of Medicine, 835 South Wolcott Avenue, Chicago, Illinois 60612. E-mail: ggibori{at}uic.edu

	Abstract

Top Abstract I. Introduction II. Formation of the... III. Genesis of a... IV. Function of the... V. Regression of the... VI. Concluding Remarks References

 
The corpus luteum (CL) is one of the few endocrine glands that forms from the remains of another organ and whose function and survival are limited in scope and time. The CL is the site of rapid remodeling, growth, differentiation, and death of cells originating from granulosa, theca, capillaries, and fibroblasts. The apparent raison d’etre of the CL is the production of progesterone, and all the structural and functional features of this gland are geared toward this end. Because of its unique importance for successful pregnancies, the mammals have evolved a complex series of checks and balances that maintains progesterone at appropriate levels throughout gestation. The formation, maintenance, regression, and steroidogenesis of the CL are among the most significant and closely regulated events in mammalian reproduction. During pregnancy, the fate of the CL depends on the interplay of ovarian, pituitary, and placental regulators. At the end of its life span, the CL undergoes a process of regression leading to its disappearance from the ovary and allowing the initiation of a new cycle. The generation of transgenic, knockout and knockin mice and the development of innovative technologies have revealed a novel role of several molecules in the reprogramming of granulosa cells into luteal cells and in the hormonal and molecular control of the function and demise of the CL. The current review highlights our knowledge on these key molecular events in rodents.

I. Introduction

II. Formation of the Corpus Luteum: Reprogramming of Follicular Cells

A. Exit from the cell cycle

B. Key molecules involved in luteinization

C. Changes in the expression of receptors

D. Activation of signaling pathways

III. Genesis of a New Gland: Structural Changes from the Remnants of the Ovulated Follicle

A. Luteal cell types

B. Tissue remodeling

C. Vascularization

IV. Function of the Corpus Luteum

A. Luteal steroidogenesis

B. Hormonal regulation

V. Regression of the Corpus Luteum: Involution to Resume Cyclicity

A. Functional luteal regression

B. Structural luteal regression

VI. Concluding Remarks

	I. Introduction

Top Abstract I. Introduction II. Formation of the... III. Genesis of a... IV. Function of the... V. Regression of the... VI. Concluding Remarks References

 
THE CORPUS LUTEUM (CL) plays a central role in the regulation of the estrous cycle and in the maintenance of pregnancy. This function is carried out largely by progesterone, which is the main steroid synthesized by this transient endocrine gland. If the oocyte is not fertilized, the CL regresses, allowing a new cycle to begin. Implantation, mating, or even cervical stimulation in some mammals initiates a complex mechanism geared to maintain CL function, ensuring a continuous supply of progesterone needed for fetal survival.

Four types of CL differing in their life span and steroidogenic output can be found in mammals, i.e., the CL of: 1) the cycle, 2) pseudopregnancy, 3) pregnancy, and 4) lactation. Only the CL of pregnancy is present in all mammalian species, whereas all four types can be found in rodents. The CL of the cycle does not exist in induced ovulators, and the CL of pseudopregnancy does not form in primates, whereas the CL of lactation is seen only in species that ovulate after parturition.

Reviews covering several aspects of the physiology of the primate CL, such as luteal steroidogenesis (1, 2), the process of luteal regression and remodeling (3), and the molecular mechanisms triggered by LH (4) have been published. The mechanism controlling luteal function, principally in ruminants, also has been reviewed by various investigators (5, 6, 7, 8, 9, 10, 11, 12). In addition, an extensive analysis of the role of immune cells and cytokines as mediators of luteal formation and regression has being published recently (7, 13, 14). Other aspects of luteal function such as angiogenesis and overall role of the luteal microvasculature have been also revisited (15, 16, 17). The clinical aspect of the CL function in assisted reproduction has also been discussed (18). Finally, an overview of the contribution of mutant mouse models to the knowledge of luteal development, function, and regression has been reported (19, 20). The present review focuses on the molecular, cellular, and physiological mechanisms underlying the processes of formation, regulation, and regression of the CL, with a particular emphasis on rodent species.

	II. Formation of the Corpus Luteum: Reprogramming of Follicular Cells

Top Abstract I. Introduction II. Formation of the... III. Genesis of a... IV. Function of the... V. Regression of the... VI. Concluding Remarks References

 
Activation of the LH receptor (LH-R) in follicular cells by the preovulatory LH surge causes ovulation and rapidly initiates a program of terminal differentiation of the ovulated follicle into a CL through a process termed luteinization. Remarkably, transformation of granulosa cells into luteal cells occurs within a few hours (21). There are structural and genomics changes that lead to the terminal differentiation of follicular cells into nondividing progesterone-producing luteal cells. Cells undergoing luteinization must stop dividing and begin expressing a new set of molecules that will allow luteal cells to survive in a different hormonal environment. Thus, the final luteal cell phenotype depends on a specific combination of genes encoding for regulatory proteins such as receptors, transcription factors, and signaling proteins, which ensures the expression of only those genes necessary for luteal cell function. This reprogramming of follicular cells into luteal cells is irreversible and requires first the exit from the cell cycle.

A. Exit from the cell cycle
LH terminates follicular growth by causing granulosa cells of preovulatory follicles to exit the cell cycle. Luteal steroidogenic cells are found arrested predominantly at the G₀/G₁ phase of the cell cycle (22). Cyclin-dependent kinases (Cdks) and several proteins that either stimulate or inhibit their activities regulate the G₁ phase of the cell cycle, governing the transition between proliferation and quiescence (23). Progression through G₁ is controlled largely by Cdks 4/6 and 2 in association with cyclins D and E, respectively. Entry of cells into S-phase involves the cooperation of Cdk 4/6 and D-type cyclins, D1, D2, and D3 with Cdk2/cyclin E. These kinases phosphorylate the retinoblastoma protein (pRb) on multiple sites. Once phosphorylated, pRb frees transcription factor E2 (E2F) that, in turn, activates diverse genes required for S-phase entry and progression (24, 25). Initiation of cell cycle arrest involves the induction of endogenous Cdk inhibitors (e.g., p21^cip1 and p27^kip1), which bind to cyclin/Cdk complexes to inhibit their activity (25). Thus, phosphorylation of pRb does not take place, allowing hypophosphorylated pRb to repress the activity of E2F leading to G₁ cell cycle arrest (reviewed in Ref. 26). Accordingly, in mice (27) and human (22) luteal cells, only unphosphorylated pRb is found and, consistent with this, they contain the Cdk inhibitor protein, p27^Kip1, but not E2F-1, which is normally expressed only in proliferating cells. More recently, the deletion of cyclin D2, p21^cip, p27^kip1, and Cdk4 genes has allowed investigators to clarify the mechanisms by which follicular cells cease to divide. Deletion of the cyclin D2 gene impairs granulosa cell proliferation and prevents follicular development. Follicles remain small, and ovulation does not occur (28). Because cyclin D2 promotes G₁ progression by activating Cdk4, a cyclin D2/Cdk4 complex was thought to be required for proliferation of follicular cells. However, Cdk4-null mice demonstrated intact follicular maturation and revealed no intrinsic irregularities in the exit from the cell cycle of follicular cells after ovulation (29, 30). Deletion of Cdk4 caused rather a rapid demise of the CL due to a profound decrease in the number of pituitary lactotropes, leading to reduced secretion of prolactin (PRL). Thus, restoring serum PRL levels is sufficient to rescue normal formation of the CL in Cdk4-null mice. The selective effect of Cdk4 deficiency on the pituitary but not the ovary suggests that other factors may compensate for its absence in the ovary. Normal cell growth in the ovary of Cdk4-null mice may depend on Cdk6 that is normally coexpressed with Cdk4 in most tissues (31) and has cyclin D-dependent kinase activity in vitro indistinguishable from that of Cdk4 (32).

Cessation of cell proliferation during luteinization is associated with a progressive loss of positive cell cycle regulators, including cyclins and Cdk2, and with increased expression of the Cdk inhibitors p21^cip1 and p27^kip1 (27, 33). Richards et al. (34) have elegantly shown that the LH surge silences, within 2 to 4 h, the expression of cyclin D2 mRNA and protein and induces that of the Cdk inhibitors p21^cip1 and p27^kip1. The expression of p21^cip1 is rapidly induced by LH; however, p27^kip1 stimulation is seen only after 12 to 24 h of treatment; this led the authors to conclude that p27^kip1 may not affect the immediate exit of granulosa cells from the cell cycle but rather contributes to the maintenance of cell cycle arrest. Deletion of p21^cip1 caused no detectable effect on proliferation of luteinized cells and fertility, whereas p27^kip1-null ovaries showed hyperproliferation of granulosa cells during luteinization, which appears to be, at least in part, the cause of the sterility seen in these mice (35). These observations suggest that p27^kip1 is a major limiting factor for successful exit of granulosa cells from the cell cycle during luteinization. The recent generation of p27^kip1, p21^cip1 double knockout (35) revealed that these two Cdk inhibitors play a cooperative role in exit of the cell cycle of granulosa cells. The absence of the two Cdk inhibitors resulted in prolonged proliferative life span of luteinized cells in vivo. This implies that exit from the cell cycle is not an obligatory step for the differentiation of granulosa cells into luteal cells. This possibility is further substantiated by the finding that granulosa cells isolated from mice lacking both Cdk inhibitors have remarkably prolonged proliferation in culture, they continue expressing granulosa cell-specific genes, yet they differentiate into luteal cells after several passages in culture despite their p27^kip1 and p21^cip1 deficiency (35). Interestingly, these cells are neither immortal nor transformed as they undergo senescence and finally die by apoptotic programmed cell death.

B. Key molecules involved in luteinization
Several genes that are rapidly and transiently induced by the LH surge are thought to be involved in ovulation and induction of luteinization. Between the transiently induced genes after the LH surge are progesterone receptor (PR) (36, 37), cyclooxygenase-2 (COX-2) (38), CATT/enhancer binding protein ß (C/EBPß) (39), early growth response protein-1 (Egr-1) (40), and Nur77 (41). Whereas Nur77-null mice are fertile, mice deficient in PR, C/EBPß, or COX-2 are infertile. This latter group of mice develops preovulatory follicles but fails to ovulate. COX-2 and PR knockout female mice do not ovulate even in response to exogenous hormones but form CL containing trapped oocytes, suggesting that luteinization can occur in the absence of these molecules (38, 42). In contrast, C/EBPß-null mice ovulate fertilizable eggs in response to gonadotropin stimulation, yet luteinization does not take place, and the CL are not formed even when the ovaries are transplanted into normal hosts (39). These data indicate that this phenotype is caused by intrinsic ovarian defect(s) and demonstrate a key role for C/EBPß in the process of luteinization. The C/EBPß transcription factor, which belongs to the basic leucine zipper class of DNA binding proteins, is rapidly induced in vivo by LH in granulosa cells (39, 43). The C/EBPß promoter contains two cAMP response elements (CREs) that may be essential for LH-stimulated C/EBPß transcription. This is because LH-R signaling is known to increase intracellular cAMP (37), which activates the CRE binding protein (CREB). The genes targeted by C/EBPß that are specifically involved in follicular cell luteinization are not yet known. Some intriguing findings were obtained with the C/EBPß-null mice, suggesting that this transcription factor does not necessarily mediate stimulation of gene expression by LH during luteinization, but rather may be responsible for silencing the expression of genes after their prior induction by LH. In wild-type mice and rats, P450 cholesterol side-chain cleavage (P450scc), P450 aromatase (P450arom), and COX-2 expressions were stimulated by LH (37, 39). However, whereas P450scc remained highly expressed long after luteinization began, both COX-2 and P450arom expressions declined several hours after LH/human chorionic gonadotropin (hCG) administration. This decrease does not occur in ovaries of C/EBPß-null mice, suggesting that a product of the C/EBPß gene could be responsible for this phenomenon. The C/EBPß gene is transcribed into a single RNA that gives rise to four C/EBPß isoforms: two full-length liver-enriched activator protein (LAP) isoforms (38 and 34 kDa), one truncated 21-kDa liver-enriched inhibitory protein (LIP) isoform, and a second truncated 14-kDa isoform. LIP C/EBPß isoforms function as dominant negative inhibitors of the full-length C/EBPß LAP isoform (reviewed in Ref. 44). Thus, functioning as a transcriptional repressor in luteinized cells, LIP may repress the expression of COX-2 and P450arom. Once luteinization occurs, C/EBPß is no longer expressed in the CL (39), suggesting that the role of this transcription factor is limited to the process of luteinization and is not involved in the maintenance of luteal function.

Other transcription factors such as Egr-1 (also named NGFI-A) and Nur77 (NGFI-B) are also stimulated by LH (40, 41, 45, 46). Both factors are encoded by immediate early genes and are capable of binding regions rich in guanine and cytosine (GC-rich) within the promoter of numerous genes. Egr-1 has been shown to be induced by LH/hCG in granulosa cells of ovulating follicles (40). Egr-1 expression in response to LH is rapid, transient, and dependent on the activation of at least two cellular signaling pathways, protein kinase A (PKA) and MAPK (46). GC-rich enhancer elements that potentially bind Egr-1 are present in many LH-regulated genes, and Egr-1 often binds overlapping sequences with Sp1, an important transcription factor controlling several ovarian-expressed genes. Based in this fact, Russell et al. (46) proposed that Egr-1 can exert either positive or negative transcriptional events by acting on Sp1 sites. Disruption of Egr-1 expression led to fertility problems due to a diminished or complete lack of expression of LHß (47). However, development of a second mutant mouse, in which the lacZ gene was introduced into the Egr-1 locus, showed that infertility results from abnormal function of both the anterior pituitary and the ovary (48). An ovarian defect is supported by the lack of ovulation and formation of CL on Egr-1 knockout mice treated with pregnant mare serum gonadotropin and hCG (48). Egr-1 knockout animals subjected to a superovulation protocol showed, however, an increase in size of both the uterus and antral follicles. These enlarged follicles lacked signs of luteinization or cumulus cells expansion; instead, the follicles appeared highly hemorrhagic. These abnormalities indicate a severe defect in the receptivity of follicular cells to LH. Accordingly, the authors demonstrated a reduction of LH-R expression in the ovary of Egr-1 knockout mice (48), suggesting that Egr-1 controls not only the expression of LHß in the pituitary but also the capacity of granulosa cells to respond to this hormone.

Classified as an orphan nuclear receptor, Nur77 displays the tripartite domain structure of members of the steroid receptor family yet does not bind any known ligand. Nur77 was shown to regulate the transcriptional activity of steroidogenic genes (49, 50, 51) and has been implicated in the regulation of apoptosis in nonovarian cell types (52, 53). Nur77 is a homolog of steroidogenic factor-1 (SF-1), which is constitutively active in many steroidogenic tissues and binds to an element similar yet distinct from Nur77 (54). In contrast to SF-1, Nur77 is present at low levels in preovulatory granulosa cells under basal conditions, but it becomes rapidly and highly expressed in response to hCG/LH stimuli both in vivo and in vitro (45). However, mice deficient in Nur77 have no apparent aberrant phenotype (reproductive or other), suggesting that biological alternative pathways may compensate for the loss of Nur77 (55). Transcription factor Nurr1 may have this function. As Nur77, Nurr1 is also an immediate-early gene with DNA and ligand-binding domains similar to that of Nur77 (56). Whether Nurr1 is expressed in the ovary in the absence of Nur77 remains to be determined.

Several members of the activator protein-1 (AP1) family of transcription factors, such as Fra2 and JunD, are stimulated during luteinization induced by LH (57), suggesting that AP1 signaling may be an important downstream target for LH. JunB, JunD, and Fra2 expression rapidly increases in response to hCG, appearing first in theca cells and thereafter in granulosa cells. JunD and Fra2, but not JunB, persist in the nucleus of luteal cells, suggesting that these factors are selectively associated with terminal differentiation of the granulosa cells (57).

C. Changes in the expression of receptors
One of the most important changes during luteinization is the alteration in the cellular responsiveness to external signals allowing luteal cells to respond to a new set of hormones. The most studied receptors are those for FSH, LH, PRL, estrogen, and progesterone. The LH surge causes the silencing of the FSH receptor (FSH-R), a transient decline in the LH-R, and a sustained stimulation of the PRL receptor (PRL-R). It also induces a rapid yet short-lived increase in PR expression and a shift in the expression of the estrogen receptor (ER) from the predominance of ERß to that of ER.

1. FSH-R.
This receptor plays an essential role during follicular development but becomes unnecessary once follicular cells differentiate. Its expression declines after the LH surge during the process of luteinization (58). This down-regulation appears to be solely due to LH because its administration to rats possessing large preovulatory follicles is sufficient to cause luteinization of granulosa cells and a marked decline in FSH-R content (59). In addition, treatment of granulosa cells with hCG completely abolished FSH-R expression (60). Once the expression of this receptor is inhibited, it does not recover and it is not expressed in the CL (60). The molecular mechanism by which LH silences the expression of the FSH-R gene is unclear, although recent investigations suggest that retinoic acid is involved in this process (61, 62). Retinoic acid, which is stimulated by LH (63), represses the FSH-R gene (62). The finding that the activity of the FSH-R promoter is repressed by retinoic acid receptor (RAR) and stimulated by SF-1, and that RAR and SF-1 bind to overlapping response elements, led Xing and Sairam (62) to propose a model for the mechanism whereby RAR and SF-1 control the overall expression of the FSH-R gene. Upon LH stimulation, RAR may displace SF-1 from the FSH-R promoter and recruit corepressors to inhibit transcription. Therefore, LH stimulation of retinoic acid may be a key step in the suppression of FSH-R during luteinization and throughout the life span of the CL. Recent findings revealed that binding of octamer transcription factor 1 to exon 1 of the FSH-R gene is required for silencing of this gene and that in Sertoli cells GATA-1 binding to the same region attenuates octamer transcription factor 1 repression (64).

2. LH-R.
The preovulatory LH surge causes first activation and thereafter desensitization of the LH-R in luteinized cells. A model for such desensitization involving ADP ribosylation factor 6 (ARF6) and arrestin 2 was recently proposed by Hunzicker-Dunn et al. (65) and shown in Fig. 1. In this model, the switch of the LH-R from an actively signaling module to one uncoupled from Gs is considered to be intimately associated with the formation of larger protein aggregates containing self-associated LH-Rs (66). A marked down-regulation of cell surface LH-R and its cognate mRNA follows desensitization (67, 68, 69). The down-regulation of LH-R mRNA that occurs under these conditions is not due to decreased transcription of the LH-R gene, but rather to increased degradation of LH-R mRNA (70). Menon and colleagues have identified and purified (71, 72) a rat ovarian protein, designated LH-R binding protein, that binds to a region of the open reading frame of the rat LH-R mRNA and enhances its degradation. In contrast to the FSH-R, the expression of the LH-R increases after luteinization and becomes highly abundant in the CL. This up-regulation of LH-R during CL formation has been shown to be due to PRL both in vivo (68, 73) and in vitro (74). However, whereas extensive investigations have defined the mechanism of LH-R down-regulation at the promoter and receptor levels (reviewed in Ref. 75), no information defining the molecular mechanism of PRL-mediated stimulation of LH-R is available to date.

View larger version (48K): [in this window] [in a new window]   FIG. 1. Model of LH-R desensitization. Shown in the top panel is the docking of arrestin 2 at a membrane location distinct from the inactive LH-R and in association with inactive ARF6 GDP. Activation of LH-R upon binding to LH, as shown in the middle panel, promotes not only activation of Gs and consequently AC activation, but also activation of ARF6 and liberation of arrestin 2. Arrestin 2 binding to the third intracellular loop of the active LH/CG-R, shown in the bottom panel, mediates desensitization by interfering with the ability of the agonist-activated receptor to activate Gs. [Adapted by permission of Federation of the European Biochemical Societies from M. Hunzicker-Dunn et al.: FEBS Letters 521:3–8, 2002 (65 ). Copyright 2002.]

The up-regulation of PRL-R during luteinization in rodents appears to depend on LH. A study reported differential stimulation of both PRL-R variants during luteinization because treatment of granulosa cells with hCG caused a 4-fold stimulation of PRL-RL and 10-fold stimulation of PRL-RS (79). Quantitative RT-PCR analysis of mRNA obtained from whole ovaries (79) revealed that levels of PRL-RL mRNA in proestrous ovaries were twice those of PRL-RS, whereas both receptor mRNAs were equally expressed in ovaries at diestrus, which contain newly formed CL. This finding supports the concept that during luteinization PRL-RS is stimulated more than PRL-RL. The molecular mechanism that leads to the increased level of one receptor mRNA over the other is not clear.

Although CL are observed in the ovaries of both wild-type and PRL-R knockout mice (80), their morphology is dramatically different. Two days after mating, wild-type mice had large CL with classical morphology, whereas PRL-R knockout mice exhibited CL undergoing regression displaying strong DNA cleavage associated with extremely low indications of vascularization. Interestingly, 1 d after ovulation, these animals have CL expressing p27^kip1 and steroidogenic enzymes at levels that are not different from the wild-type animals. The major defect seen in the CL of the PRL-R-knockout mice is the premature expression of the 20-hydroxysteroid dehydrogenase (20HSD) enzyme that is present as early as the first day of pregnancy. This enzyme is known to catabolize progesterone to 20-dihydroprogesterone (20DHP), and its expression may be in large part responsible for the lack of implantation observed in the PRL-R knockout mice. Two days after ovulation, the CL of these animals shows an elevated degree of apoptosis together with decreased levels of p27^kip1 and steroidogenic enzymes (80). These data support the importance of PRL and progesterone as antiapoptotic hormones needed for luteal survival.

4. ER.
In rodents, the ovary expresses both ER and ERß genes as single (6.5 kb) and multiple transcripts (ranging from 1.0 kb to approximately 10 kb), respectively (81, 82). Whereas ERß is abundantly expressed in the follicle, especially in the granulosa cell layer (81, 82), ER is the major receptor found in the CL (83) at levels 10-fold higher than that of ERß (Fig. 2). In granulosa cells, a large decline in ERß expression occurs during proestrous. This down-regulation is clearly associated with the LH surge and can be mimicked in vivo and in vitro by hCG stimulation (81, 84). Moreover, agents stimulating LH/hCG receptor-associated intracellular signaling pathways (e.g., forskolin and a phorbol ester) readily mimic the effect of hCG in down-regulating ERß mRNA in cultured granulosa cells (81). LH-induced down-regulation of ERß mRNA levels is not due to an effect on transcription, but rather on mRNA stability. This LH-induced destabilization of the ERß mRNA requires ongoing protein synthesis (85). Although the molecular mechanisms involved in LH-induced destabilization of ERß mRNA await further studies, it may be attributable to the blockade of translation elongation of ERß transcripts themselves or to the interaction between LH-induced new protein(s) and the ERß mRNA, as suggested by Park-Sarge and co-workers (85). These LH-induced protein(s) may alter the tertiary folding of the ERß transcripts, rendering them easy targets for ribonucleases. It is not yet known, however, whether the decrease in ERß during luteinization is essential for the normal development of the CL. It has been shown that ERß can act as ER-dominant negative and repress its transcriptional activity, leading to an overall decrease in the cellular sensitivity to estradiol (86). Because both ER and ERß remain expressed throughout the life span of the CL of the pregnant rodent, mainly because of PRL stimulation (82, 83), a function for both luteal ERs throughout pregnancy remains a possibility.

View larger version (35K): [in this window] [in a new window]   FIG. 2. Expression of ER and ERß mRNAs in the pregnant rat ovary as shown by in situ hybridization. Left panel shows abundant expression of ERß mRNA mainly at the small growing follicles, whereas the right panel shows abundant expression of ER mRNA mainly in the CL of pregnancy. The ovaries were from d 17 pregnant rats. [Adapted from C. M. Telleria et al.: Endocrinology 139:2432–2442, 1998 (82 ). Copyright The Endocrine Society.]

D. Activation of signaling pathways
Differentiation to a luteal phenotype also implies numerous changes in the intracellular signaling network activated upon receptor occupancy. Several signaling pathways participate in the profound changes that take place during luteinization and CL formation. See Fig. 3 for a scheme of the multiple factors affected by the surge of LH and their participation in the different processes induced by this hormone.

View larger version (58K): [in this window] [in a new window]   FIG. 3. Transient and permanent effects of LH on key follicular genes leading to luteinization. See text for details.

MAPKs have been implicated in the phosphorylation of CREB. MAPKs comprise a superfamily of kinases (99) that have been identified in the CL of several species. ERK1 and ERK2 are expressed in porcine, bovine, rat, and human CL (100, 101, 102, 103, 104, 105), and p38/MAPK has been shown to be expressed in the CL of rat and cow (101, 106), whereas Jun N-terminal kinase has been studied mostly in bovine CL (106, 107). Hunzicker-Dunn and colleagues (101) have elegantly demonstrated that activation of LH-R leads to phosphorylation of p38/MAPK and its upstream activator MAPK kinase 6 but has no effect on either MAPK kinase 3 or ERK1/2. The p38/MAPK downstream protein kinase target termed MAPK-activated protein kinase 3 (MAPKAPK3) is induced at both mRNA and protein levels during luteal formation, whereas mRNA and protein expression of the closely related MAPKAPK2 is diminished. MAPKAPK3, activated during luteinization in vivo, readily catalyzed CREB phosphorylation in immune complex kinase assays, and phosphorylation of CREB was shown to depend on an intact p38/MAPK signaling pathway in a cellular model of luteinization. Thus, activation of the MAPKAPK3 seems to be crucial to maintain CREB phosphorylation during luteinization (101). Whether LH activation of this MAPK system is mediated by the increase on cAMP or by growth factors such as IGF-I or vascular endothelial growth factor (VEGF) needs to be investigated.

2. LH/PLC/PKC.
The LH surge not only generates high levels of intracellular cAMP in granulosa cells, but also activates phospholipase C (PLC), leading to the increase in intracellular calcium (108, 109, 110). The actions of ovulatory concentrations of LH can be mimicked by subovulatory doses of LH/hCG plus a protein kinase C (PKC) activator in primary granulosa cell cultures (111, 112), suggesting that the effects of LH are mediated by more than one kinase. However, despite the ability of the LH-R to activate PLC under experimental conditions and the ability of PKC activators to synergize with hCG, the responses to LH appear to be independent from the activation of PKC (113).

3. LH/Wnts/Frizzled.
Wnt proteins are secreted extracellular signaling molecules that act locally to control diverse developmental processes such as proliferation and differentiation. Wnts transduce their signals by binding to G protein-coupled receptors of the Frizzled (Fzd) family (for review, see Ref. 114). In the rodent’s ovary, Wnt-4 performs critical functions during early ovarian development (115). Wnt-4 and Wnt-2 mRNA are expressed in granulosa cells (116, 117). Wnt-4 expression increases after hCG treatment and remains elevated in the CL during pregnancy (117). Luteal cells express Disheveled 1 (Dvl-1) and ß-catenin (117), which are components of Wnt-4 intracellular signaling pathway. Expression of Fzd1 and Fzd4 is also stimulated by LH-R activation. Although the role of both Wnt-4 and Fzd4 in luteinization is not yet clear, the results obtained to date suggest that Wnt-4 may be a ligand for Fzd4 and that Wnt-4/Fzd4 signaling is important for the regulation of luteal cell formation and function. The recent deletion of the Fzd4 gene has revealed that the null mice fail to form functional CL despite normal follicular development and ovulation of fertilizable oocytes. Fzd4–/– ovaries exhibit CL of altered appearance and reduced expression of genes known to be associated with luteinization (118). Because these genes are regulated by PRL and because the phenotype of the Fzd4–/– null mice is similar to that of PRL-R–/– mice, Hsieh et al. (118) speculate that PRL signaling might be defective in the newly formed CL of Fzd4–/– mice.

4. PRL/Jak/Stat.
This pathway, formed by the Janus kinase (Jak) and the signal transducers and activators of transcription (Stat) proteins, is activated after ovulation mainly by PRL. PRL binding to the long form of its cognate receptor (PRL-RL) was thought to induce receptor homodimerization, which leads to immediate transphosphorylation of the associated tyrosine kinase Jak2, followed by phosphorylation of Stat transcription factors Stat5a and Stat5b (119). Recent data indicate, however, ligand-independent dimerization of the PRL-R (120). The increase in PRL-R expression observed during luteinization is accompanied by changes in the activation of the intracellular signaling transduction pathway of PRL (34, 121, 122). The activity of Stat5a and Stat5b is nondetectable in preovulatory follicles (79), but it is induced by PRL after the LH surge (121). PRL activates Stat5b as well as Stat5a; however, Stat5b is the predominant isoform in luteal cells (79). Recent investigations have revealed that there is a difference in the mechanism by which PRL activates Stat5a and Stat5b (123). Whereas PRL activation of Jak2 leads to phosphorylation and activation of Stat5a, Jak2-induced tyrosine phosphorylation of Stat5b can be separated from Stat5b transcriptional activity. PRL appears to activate a second tyrosine kinase that can phosphorylate Stat5b at a place that does not lead to Stat5b transcriptional activity. The high sequence similarity between Stat5a and Stat5b (96%) suggests that there may be redundancy in their signaling pathways, whereas differences in their tissue distribution may lead to distinct functions. Mice knockout for Stat5a do not have major reproductive defects; they suffer only from impaired mammary gland development (124). Stat5b, on the other hand, seems to be essential for luteal function and survival because Stat5b-deficient mice abort beyond d 7 of pregnancy, a phenomenon that could be partially prevented by treatment with progesterone (125). Because abortion occurs after d 7 of pregnancy, it appears that the major role of Stat5b is not in luteinization, but rather in CL maintenance and production of progesterone. Interestingly, only the double knockout Stat5a/b female mice, but not the Stat5a or Stat5b mutants, are totally infertile, demonstrating a functional redundancy of the Stat5 proteins (126). Stat5a/b double mutant mice have normal developing follicles and ovulation, yet few CL are evident in their ovaries, indicating that Stat5 molecules may play a crucial role in the development and survival of the gland (126). However, because the luteal phenotype of Stat5a/b null mice has not been investigated with sufficient detail, it is not clear whether CL formation and luteinization, which occur rapidly after ovulation, are affected by the lack of the Stats. As in the case of the PRL-R null mice, disruption of the Stat5a/b genes results in low levels of p27^kip1, leading to the speculation that the Stat5 proteins control luteal p27^kip1 expression and that this expression is critical for inducing differentiation of follicular cells into luteal cells. Of great interest is the finding that Stat5a/b deletion leads to extensive expression of the progesterone metabolizing enzyme 20HSD in their CL (126), thus explaining the low circulating levels of progesterone found in these animals (see Section IV). As mentioned earlier, lack of Stat5b is associated with early abortions that are due to reduced progesterone levels during midgestation (125). This deficiency is partially corrected in the double knockout for 20HSD and Stat5b (127), supporting the concept that activation of Stat5b is important in suppressing 20HSD gene expression.

Deactivation of the Jak/Stat signaling pathway plays an important role in establishing and maintaining tissue responsiveness to PRL. Negative regulation of cytokine signaling downstream of Stat activation occurs through several families of modulators. Two Src-homology domain-containing protein tyrosine phosphatases (SHP-1 and SHP-2) have been shown to modulate PRL-induced Stat responses. SHP-2 is a substrate for Jak2 and is obligatory for PRL/Stat5 induction of ß-casein (128). On the other hand, SHP-1 binds activated Stat5b, catalyzing its rapid dephosphorylation, leading to the transient activation pattern that is characteristic of this family of transcription factors (129). SHP-1 expression (mRNA and protein) is undetectable in preovulatory follicles, but it is maximally expressed in the nuclei of luteal cells throughout gestation (79), suggesting acquisition of PRL responsiveness and activation of Stat5b. The presence of SHP-1 in the nucleus of luteal cells explains the transient Stat5b activation in PRL-treated CL during early gestation. These results led Russell and Richards (79) to postulate that, at least in rodents, SHP-1 may continuously dephosphorylate activated Stat5b, resulting in a transient activation profile during pulsatile PRL exposure early in pregnancy.

5. PI3K/Sgk/Foxo.
The forkhead family of transcription factors, FKHR (Foxo1), FKHRL1 (Foxo3), and AFX (Foxo4) are regulated by a pathway involving phosphatidylinositol-3-kinase (PI3K) and protein kinase B (PKB/Akt). Direct phosphorylation by PKB or by the PKB-related kinase, serum and glucocorticoid-induced kinase (Sgk), results in translocation to the cytoplasm and inactivation of the forkhead transcription factors (for review, see Ref. 130). As granulosa cells luteinize, they gain increased levels of Sgk, Foxo4, and Foxo3, whereas at the same time, Foxo1 expression declines, possibly affecting the expression of p27^kip1 and p21^cip1 (131). Deletion of Foxo3a led to premature ovarian failure (132), therefore establishing a key role for this transcription factor in the survival of granulosa cells and oocytes.

6. Oocyte/TGFß pathway.
It has long been believed that oocyte-derived regulatory molecules act within the follicle to inhibit premature luteinization and limit progesterone biosynthesis (133). In vitro studies have demonstrated that members of the TGFß superfamily mediate the antiluteinizing effect of the oocyte (134, 135). The TGFß superfamily includes TGFß, activins and inhibins, bone morphogenetic proteins (BMPs), growth/differentiation factors (GDFs), and anti-Müllerian hormone (136). BMP-15 (134) and GDF-9 (135), both produced by the oocyte, have been proposed as luteinization inhibitors. TGFß-related proteins signal through a serine–threonine kinase cascade that results in the cytoplasmic to nuclear translocation of intracellular effector proteins termed "mothers against decapentaplegic homolog (SMADs)." TGFß and activin signal through SMAD2 and SMAD3, whereas the BMPs and GDFs mediate their signals through SMAD1, SMAD5, and SMAD8 (for review, see Ref. 137). In addition, the capacity for the TGFß-related ligands to signal is thought to require SMAD4, the common SMAD that partners with the receptor regulated-SMADs to form the core of a transcriptional complex (138). SMAD4, therefore, is a central component of the TGFß superfamily signaling pathway. Recent generation of SMAD4 and activin granulosa cells conditional knockout resulted in premature luteinization and ovaries full of CL, respectively. This provides clear in vivo evidence that activin prevents premature luteinization of granulosa cells during follicle development (139). Interestingly, these changes are accompanied by a significant increase in serum progesterone levels (139) and high levels of luteal markers such as P450scc, 17ßHSD-7, and steroidogenic acute regulatory protein (StAR).

	III. Genesis of a New Gland: Structural Changes from the Remnants of the Ovulated Follicle

Top Abstract I. Introduction II. Formation of the... III. Genesis of a... IV. Function of the... V. Regression of the... VI. Concluding Remarks References

 
The formation of the CL not only involves changes in the expression of genes implicated in the cell cycle and signaling in theca and granulosa cells but also the breakdown of the follicular basal membrane necessary for rapid migration of endothelial cells, fibroblasts, and theca cells into the previously avascular granulosa layer. The CL is therefore the product of luteinization of follicular cells, endothelial cell invasion, and tissue remodeling.

A. Luteal cell types
The CL is a heterogeneous gland composed of small and large steroidogenic luteal cells, fibroblasts, endothelial, pericytes, and immune cells. These cells have different morphological, endocrine, and biochemical features. Interactions between these various cell types is essential in maintaining the health and steroidogenic function of the CL (16, 140, 141, 142). Once luteinization takes place, a defined cell population undergoes extensive hypertrophy and differentiates into large steroidogenic luteal cells, whereas another cell population remains much smaller and comprises small steroidogenic luteal cells. In ruminants and rodents, small and large luteal cells differ in their basal rates of progesterone secretion, with the large cells producing 2- to 40-fold more progesterone than the small cells. These two luteal cell types, however, differ in their response to different hormonal and/or second messenger stimuli (140). It is widely believed that the origin of the large luteal cells is the granulosa cells, whereas the theca cells differentiate into small luteal cells; however, although this hypothesis is well supported in domestic animals (for review, see Ref. 6), there is no evidence indicating these cellular origins in rodents (140, 143). In most species there is considerable mixing of large and small cell types during the reorganization of the follicle into the CL leading to a close contact between the two cell types. Primates are an exception, with the two cell populations remaining relatively separate, and therefore are generally called granulosa-lutein cells and theca-lutein cells, respectively (6). These two cell populations can be distinguished in tissue sections by their location; furthermore, whereas theca-lutein cells do not hypertrophy during the luteal phase, the granulosa-lutein cells undergo considerable hypertrophy that significantly contributes to the growth of the CL (144, 145). Additionally in primates, theca-luteal cells are the primary source of androgens (146), whereas granulosa-luteal cells are the site for estrogen synthesis (147), indicating that the two-cells model of estrogen biosynthesis invoked to explain follicular estrogen production is preserved in the CL of primates.

In contrast to the primate CL, rodent’s small and large luteal cells are intermingled within the CL and do not appear to differentiate from the theca and granulosa cell, respectively. The protein profile of both cell types is largely similar (Fig. 4) with the only major difference being the abundant expression of a 32-kDa protein known as PRAP (for PRL-R associated protein) (148, 149). PRAP is found solely in large luteal cells and is an excellent marker for them. Yet in the follicle, it is the theca cell layer and not the granulosa cell layer that expresses this protein (see Section IV and Fig. 8). This, added to the finding that both the large and small luteal cells express P450 17-hydroxylase/C17–20-lyase and cytochrome P450 aromatase (140, 148, 150) and produce both androgen and estrogen, casts serious doubt to the dual origin of both luteal cell types.

View larger version (29K): [in this window] [in a new window]   FIG. 4. Expression of PRAP in large luteal cells. Large and small luteal cells were isolated from rat CL, and proteins were separated by gel electrophoresis. [Adapted from M. P. McLean et al.: Endocrinology 126:1796–1805, 1990 (148 ). Copyright The Endocrine Society.]

View larger version (34K): [in this window] [in a new window]   FIG. 8. Expression of 17ßHSD-1 and 17ßHSD-7 in ovaries of pregnant rats and the reactions that they catalyze in the rat CL. 17ßHSD-1 (white signal) is expressed only in granulosa cells (from Ref. 274 ). In contrast, 17ßHSD-7 (red staining) is expressed in the CL and the theca (T) and interstitial (I) cells, but not at all in the granulosa cells (C. T. Albarracin and G. Gibori, unpublished observations). Lower left panel, In the follicle, 17ßHSD-1 has dual activity and can convert androstenedione to testosterone as well as estrone to estradiol. Lower right panel, After the LH surge, 17ßHSD-1 is no longer found in the CL, which expresses instead PRAP/17ßHSD-7, an enzyme that only converts estrone to estradiol.

C. Vascularization
The development of capillaries from preexisting blood vessels is essential for the formation and function of the CL (15, 161). Each luteal cell is in direct contact with several capillaries, giving the CL one of the highest rates of blood flow in the organism. The development of a new microcirculatory bed involves ECM degradation, endothelial cell proliferation, expansion of the capillaries, and development of capillary lumen (maturation). The dense capillary network that is formed efficiently supplies nutrients, hormones, and lipoprotein-bound cholesterol to the luteal cells and provides a mechanism for speedy and efficient output of progesterone from the CL. Soon after the LH surge and ovulation, pericytes derived from the theca compartment are the first vascular cells to invade the developing luteal parenchyma. These pericytes rapidly proliferate and populate a large percentage of vessels in the mature CL (162). In the rat CL, collagen type IV and laminin are detected in the granulosa layer approximately 6 h after ovulation, whereas within 16 h, new complete capillaries can be found (163, 164). The molecular regulation of angiogenesis in the CL is complex, with a growing list of regulators including VEGF that are essential for CL angiogenesis (165), basic fibroblast growth factor (bFGF), the newly discovered endocrine gland-derived VEGF (EG-VEGF) (166), and angiopoietins (Ang). LH /hCG appears to stimulate the expression of VEGF, EG-VEGF, Ang, and their receptor and to affect profoundly vascularization of the CL during luteinization (167) (see Fig. 5).

View larger version (34K): [in this window] [in a new window]   FIG. 5. Stages of luteal vascularization and their control by LH. See text for details.

Recently, a novel regulator of ovarian angiogenesis, EG-VEGF or Prokineticin-1 (PK-1), was identified in the ovary and has been proposed as a steroidogenic gland-specific angiogenic regulator. In contrast to VEGF, in human and nonhuman primates, EG-VEGF expression is largely restricted to the steroidogenic glands such as the ovary, testis, adrenal cortex, and placenta (179). Consistent with such an expression pattern, the human EG-VEGF gene promoter has a potential binding site for SF-1, a pivotal element for steroidogenic-specific transcription. Although structurally different, EG-VEGF and VEGF have similar functions; both promote proliferation, survival, and chemotaxis of endothelial cells isolated from steroidogenic tissues (166, 180). G protein-coupled receptors of the neuropeptide Y receptor class have been identified in endothelial cells as the cognate receptors for EG-VEGF (181). The mitogenic and prosurvival activities of EG-VEGF correlate with the ability of this peptide to induce phosphorylation of ERK1/2 and Akt (181). In humans, peak VEGF expression is found in early luteal phase, associated with the initial development of a capillary plexus within the CL, whereas EG-VEGF is low in early stage of the CL but is up-regulated during midluteal to late luteal phase at a time when VEGF expression is considerably reduced (167). When delivered by adenovirus in mouse, EG-VEGF induced extensive angiogenesis in the ovary but not in other tissues (179). Because EG-VEGF and VEGF have an additive response in vitro (166), they may also cooperate in vivo to induce the fenestrated phenotype and to promote angiogenesis, especially because both molecules are produced in the CL in a complementary and coordinated fashion. Taken together, these observations support the notion that VEGF activity is rate limiting for the creation of the capillary plexus within the CL, whereas EG-VEGF may stimulate, together with VEGF, the angiogenesis that accompanies early-to-mid CL development.

The first angiogenic factor identified in the ovary was bFGF, (182) which is produced by steroidogenic and endothelial cells of rat (183), human (184), and ruminant CL (185). bFGFs have been found to stimulate proliferation and motility of luteal endothelial cells (185, 186, 187), and treatment with antibodies to bFGF suppressed approximately 80% of endothelial cell proliferative activity in extracts of CL from cows, sheep, and pigs (185). Surprisingly, however, deletion of the bFGF gene in mice did not result in developmental disruption or loss of fertility (188), raising doubts as to how crucial bFGF is for luteal angiogenesis.

Ang belong to a family of growth factors that appear to be also critical for angiogenesis and vessel integrity. They have been implicated in the regulation of vascular development and regression. Ang-1 and Ang-2 bind the Tie-2 receptor; however, whereas Ang-1 stimulates sprouting and maturation of blood vessels, Ang-2 inhibits this effect, acting as a competitive inhibitor by binding the Tie-2 receptor without activating intracellular signaling pathways (for review, see Ref. 189). Tie-2 or Ang-1 knockout mice are embryonic lethal, with the most prominent defects involving the vasculature (190, 191). Ultrastructural analysis of the blood vessels in these animals reveals defects in endothelial cell interactions with their basement membranes and a marked decrease in the number of periendothelial support cells. Little information on Ang and their receptors is available in the CL despite the specificity of the site of their expression in rodents and primates and their apparent regulation by LH/hCG (192, 193). Ang-2 transcripts were found in close association with blood vessels in the theca interna of the preovulatory follicle and in the front of vessels invading the developing CL. Ang-1 transcripts were also associated with blood vessels but appeared to follow rather than to precede vessel ingrowth into the early CL. These expression patterns have led Maisonpierre et al. (192) to suggest that Ang-2 may collaborate with VEGF at the front of invading vascular sprouts by blocking the function of Ang-1, whereas Ang-1 may have a later role than VEGF in angiogenesis involving vessel maturation and/or stabilization.

	IV. Function of the Corpus Luteum

Top Abstract I. Introduction II. Formation of the... III. Genesis of a... IV. Function of the... V. Regression of the... VI. Concluding Remarks References

 
A. Luteal steroidogenesis
Severe changes in the expression of steroidogenic enzymes and in the type of steroid produced occur during and after luteinization. In rodents, luteal cells continue to synthesize androstenedione and estradiol, but become a substantial site of progesterone biosynthesis. This is in contrast to the follicle, where progesterone mainly serves as substrate for estradiol production. The CL expresses high levels of key proteins involved in the uptake, synthesis, and transport of cholesterol, and in the processing of cholesterol to progesterone, androgens, as well as estrogens. These proteins are the sites of regulation by tropic hormones such as PRL, LH, and estradiol (see Fig. 6 for summary).

View larger version (29K): [in this window] [in a new window]   FIG. 6. Diagram depicting genes involved in synthesis of progesterone, 20-dh-progesterone (20-DHP), and estradiol and their regulation by different tropic hormones. +, Stimulatory effect; –, inhibitory effect.

View larger version (48K): [in this window] [in a new window]   FIG. 7. Oil-red staining of lipids in ovaries from SR-BI+/+ or SR-BI–/– mice. [Adapted from B. Trigatti et al.: Proc Natl Acad Sci USA 96:9322–9327, 1999 (213 ).]

Once cholesterol reaches the outer mitochondrial membrane, it is transported to the inner mitochondrial membrane through the aqueous intermembrane space. This step involves several proteins, including the StAR (for review, see Refs. 224 and 225), peripheral-type benzodiazepine receptor (PBR) (226, 227), and possibly hormone-sensitive lipase (228). The role of StAR in the ovary was further substantiated by the phenotype of the StAR-null mice (229). Immediately after birth, the ovaries of these mice appear normal. After puberty, however, lipid accumulation occurs, and this is accompanied by incomplete follicular maturation that ultimately ends in premature ovarian failure. In contrast to the StAR-null ovaries, no detectable lipid deposits were present in the interstitial and theca cells of double StAR- and gonadotropin-knockout mice, suggesting that gonadotropin stimulation is essential for lipid accumulation in the ovaries of StAR-deficient mice (230). The impaired ovulation of the double StAR/gonadotropin-null mice cannot be reversed with exogenous progesterone and gonadotropin, suggesting that the compromised steroidogenesis of the StAR-deficient ovaries makes ovulation impossible. The mechanism whereby StAR facilitates cholesterol transport has been under intense investigation. It was originally thought that StAR might act as an intermitochondrial shuttle. However, restriction of StAR to the cytoplasm does not inhibit its activity, whereas importation of StAR into the mitochondria causes its degradation (231), suggesting that StAR acts on the outer surface of the mitochondria.

The expression of StAR in the CL was shown to be up-regulated by LH/cAMP (232) and to be repressed by prostaglandin F2 (PGF2) (233). Transcriptional regulation of the StAR gene is the primary mechanism of control of steroidogenesis and has been the focus of many studies. Comparison of StAR promoters across species indicates that the first 250 bases of the proximal promoter are critical for basal and hormone-stimulated StAR transcription and that the transcription factor response elements are highly conserved in this region (234). Several transcription factors, including SF-1, C/EBPß, SREBP-1a, cFos, GATA-4, Sp-1, and CREB family members have been implicated in the transcriptional stimulation of the StAR gene (for review, see Ref. 235). Conversely, dosage-sensitive sex reversal-adrenal hypoplasia congenital critical region on the X chromosome gene 1 (DAX-1) and forkhead box protein L2 (FOXL2) were shown to play key roles on its repression (236).

The PBR, on the other hand, is a high-affinity, cholesterol-binding protein found in between outer and inner mitochondrial membranes (237). In this location, PBR could function as a pore, allowing for the translocation of cholesterol from the outer to the inner mitochondrial membrane (238). PBR is closely associated with StAR at the outer mitochondrial membrane (239) and appears to be a promising partner for it. The importance of these two proteins in luteal steroidogenesis is suggested from the studies of Sridaran et al. (240, 241). These investigators have demonstrated that PBR and StAR are coexpressed in the pregnant rat CL and that their coordinated suppression by GnRH agonists leads to reduced progesterone production followed by luteal cell death (242). Recently, the hormone-sensitive lipase that is responsible for the neutral cholesteryl ester hydrolase activity in steroidogenic tissues was shown to interact with StAR and to facilitate cholesterol movement from lipid droplets to the mitochondria for steroidogenesis (228).

3. Biosynthesis of progesterone by the CL.
Once cholesterol reaches the inner mitochondrial membrane, its transformation into steroid hormones begins. The capacity to transform cholesterol to progesterone is a universal characteristic of the CL and involves the mitochondrial P450scc and one of the six isoforms of the 3ßHSD, 3ßHSD type II (3ßHSD-II), which is located in the smooth endoplasmic reticulum (243). Formation of the CL is accompanied by a dramatic increase in the expression of both enzymes (214, 244, 245) and in the organelles that house them (246, 247), allowing the CL to synthesize large amounts of progesterone. Both enzymes remain highly expressed in the CL throughout pregnancy and are considered to be constitutively expressed (245, 248), yet various investigators have shown that they can be regulated by PRL and gonadotropin (245, 249, 250, 251, 252, 253). In the case of 3ßHSD-II, PRL stimulation is at the level of transcription and is mediated by transcription factor Stat5 (254).

The level of progesterone secreted by the CL in rodents depends not only on the amount of progesterone synthesized by the luteal cells but also on the expression of the enzyme 20HSD that catabolizes progesterone into the inactive progestin, 20-DHP. Once 20HSD becomes expressed in the CL, progesterone secretion drops and 20-DHP becomes the major steroid secreted by luteal cells (255, 256). Due to the detrimental effect of 20HSD on luteal progesterone secretion, its pattern of expression/activity greatly contributes to the changes in the circulating levels of progesterone during gestation, ultimately controlling the progression of pregnancy. Indeed, it has been known for years that little, if any, 20HSD activity is found in the rat CL throughout pregnancy; however, elevated activity is found just before parturition concomitant with the rapid decrease in the concentration of progesterone in serum. In rodents, this decrease in serum progesterone is essential for parturition to take place. Accordingly, mice deficient for 20HSD sustain high progesterone levels and display a delay in parturition for several days (127). The development of a specific antibody against rat 20HSD (257) and the cloning of its gene (258, 259) led to the demonstration that the lack of enzyme activity throughout pregnancy and its appearance just before parturition are not due to activation/deactivation of an already present enzyme but rather to changes in gene expression (255, 257, 260). The crystalline structures of human and rabbit 20HSD have been recently obtained (261). Although 20HSD is one of the most important players in the regulation of luteal progesterone secretion in rodents, the role of this enzyme in the control of progesterone production and/or action in other species needs further investigation.

Two other enzymes have been implicated in the reduction of progesterone secretion by the CL at the end of pregnancy. P450c26, which catalyzes the conversion of cholesterol to 26-hydroxycholesterol, also rises at the end of pregnancy and could reduce progesterone secretion by limiting substrate availability (262). Luteal 5-reductase, which converts testosterone to dihydrotestosterone (DHT), has also been implicated in the functional demise of the CL. This is supported by the secretion of DHT from the rat CL at the end of pregnancy and by the fact that exogenous DHT can decrease progesterone biosynthesis (263, 264). However, although deletion of the 5-reductase type I gene impairs cervical ripening, causing parturition defects, it does not affect the decline in serum progesterone concentration at the end of pregnancy (265), raising doubts about the physiological role of DHT during luteal regression in rodents.

4. Biosynthesis of androgen and estradiol by the CL.
In rodents as in humans, the CL also produces androgens and estrogens in addition to progesterone. The major androgen produced by the ovary is the weak androgen, androstenedione (266). Conversion of progesterone to androstenedione is mediated by the enzyme P45017-hydroxylase/C17–20 lyase (P450c17 or CYP17). In preovulatory follicles, P450c17 is expressed in the theca interna and interstitial cells but not in granulosa cells (267), whereas in the CL both luteal cell types express P450c17 (140) and synthesize aroma