This Article Abstract Full Text (PDF) An erratum has been published 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 Palter, S. F. Articles by Adashi, E. Y. Search for Related Content PubMed PubMed Citation Articles by Palter, S. F. Articles by Adashi, E. Y.

Are Estrogens of Import to Primate/Human Ovarian Folliculogenesis?¹

Division of Reproductive Endocrinology (S.F.P.), Department of Obstetrics and Gynecology, Yale University School of Medicine, New Haven, Connecticut; Division of Reproductive Sciences (A.B.T., E.Y.A.), Department of Obstetrics and Gynecology, University of Utah Health Sciences Center, Salt Lake City, Utah; Department of Obstetrics and Gynecology (A.H.), Sheba Medical Center, Tel Aviv University, Sackler School of Medicine, Tel-Hashomer, Ramat-Gan, Israel; and Division of Endocrinology and Metabolism (J.D.V.), University of Virginia Health Sciences Center, Charlottesville, Virginia

	Abstract

Top Abstract I. Introduction II. The Nonprimate Ovary... III. The Nonprimate Ovary... IV. Lessons Learned from... V. Lessons Learned from... VI. The Primate/Human Ovary... VII. Is an Estrogen-Free... VIII. The Primate/Human Ovary... IX. Estrogen Reception and... X. Summary XI. Directions for Future... References

 
The notion that estrogens play a meaningful role in ovarian folliculogenesis stems from a large body of in vitro and in vivo experiments carried out in certain rodent models, (e.g., rats) wherein the stimulatory role of estrogen on granulosa cell growth and differentiation is undisputed. However, evidence derived from these polyovulatory species may not be readily generalizable to the monoovulatory subhuman primates, let alone the human. Only recently, significant observations on the ovarian role(s) of estrogen have been reported for the primate/human. It is thus the objective of this communication to review the evidence for and against a role for estrogens in primate/human ovarian follicular development with an emphasis toward the application of the concepts so developed to contemporary reproductive physiology and to the practice of reproductive medicine. The role(s) of estrogens will be examined not only by analyzing the physiological evidence to the effect that these hormones control ovarian function and follicular growth, but also by summarizing the molecular evidence for the existence and distribution of the cognate receptors.

I. Introduction

II. The Nonprimate Ovary as a Site of Estrogen Reception

III. The Nonprimate Ovary as a Site of Estrogen Action

A. The rat and mouse

B. The hamster

C. The rabbit and guinea pig

D. The pig

E. Possible interactions of estrogen with other putative ovarian regulators

IV. Lessons Learned from the Estrogen-Resistant Model- Estrogen Receptor Null Mutants (ERKOs)

V. Lessons Learned from the Targeted Disruption of the CYP-19 (Aromatase) Gene-Aromatase Null Mutant (ArKO)

VI. The Primate/Human Ovary as a Site of Estrogen Reception

A. Molecular probing

B. Immunohistochemical studies

VII. Is an Estrogen-Free (or at Least Poor) Intrafollicular Environment Compatible with Follicular Development, Ovulation, and Corpus Luteum Formation?

A. Follicular "expansion" vs. follicular "growth": an important conceptual distinction

B. Lessons learned from 17-hydroxylase/17–20 lyase deficiency

C. Lessons learned from the 3ß-hydroxysteroid dehydrogenase (3ß-HSD) deficiency

D. Lessons learned from aromatase deficiency

E. Lessons learned from the intensely hypogonadotropic model

F. Impact of estrogen deficiency on oocytic and early embryonic development

VIII. The Primate/Human Ovary as a site of Estrogen Action

IX. Estrogen Reception and Action: The Nonclassical Alternative(s)

X. Summary

XI. Directions for Future Research

	I. Introduction

Top Abstract I. Introduction II. The Nonprimate Ovary... III. The Nonprimate Ovary... IV. Lessons Learned from... V. Lessons Learned from... VI. The Primate/Human Ovary... VII. Is an Estrogen-Free... VIII. The Primate/Human Ovary... IX. Estrogen Reception and... X. Summary XI. Directions for Future... References

 
THE NOTION that estrogens play a meaningful role in ovarian folliculogenesis stems from a large body of in vitro and in vivo experiments carried out in certain rodent models (e.g., rats) wherein the stimulatory role of estrogen on granulosa cell growth and differentiation is undisputed (1, 2). However, evidence derived from these polyovulatory species may not be readily generalizable to the monoovulatory subhuman primates, let alone the human. Only recently, significant observations on the ovarian role(s) of estrogen have been reported for the primate/human. It is thus the objective of this communication to review the evidence for and against a role for estrogens in primate/human ovarian follicular development with an emphasis toward the application of the concepts so developed to contemporary reproductive physiology and to the practice of reproductive medicine.

In this regard, the evidence relevant to the human will be compared to that of other species, to be presented in the first part of this review. The role(s) of estrogens will be examined not only by analyzing the physiological evidence to the effect that these hormones control ovarian cell function and follicular growth, but also by summarizing the molecular evidence for the existence and distribution of the cognate receptors. Based on cumulative data in several species, an operational model appears to be emerging, in large part due to the discovery of a new estrogen receptor (ER) type (i.e., ERß), the identification of human examples of specific mutations in enzymes of the steroidogenic pathway, as well as the generation of specific null mice mutants for the ER, ERß, and the aromatase genes.

It must be pointed out, however, that even though this review will question the role of estrogens in primate/human follicular development, one must remain cognizant of the unassailable extraovarian roles of estrogens in these species. Clearly, estrogens play a vital role in supporting the growth and differentiation of the Mullerian (3) and mammary (4, 5) complexes. Of equal import is the role played by estrogens in the synchronization of ovarian follicular development with the midcycle gonadotropin surge (6) and the preparation of the uterus for implantation (7, 8, 9). Somewhat less well characterized are the "nonreproductive" effects of estrogenic substances, which may include effects on the cardiovascular (10, 11) and skeletal (12, 13) systems. Intriguing research has also been conducted to describe the role of estrogens in the central nervous system wherein it appears that estrogens may play a physiological role in the modulation of mood, cognition, and behavior (14, 15).

	II. The Nonprimate Ovary as a Site of Estrogen Reception (Table 1)

Top Abstract I. Introduction II. The Nonprimate Ovary... III. The Nonprimate Ovary... IV. Lessons Learned from... V. Lessons Learned from... VI. The Primate/Human Ovary... VII. Is an Estrogen-Free... VIII. The Primate/Human Ovary... IX. Estrogen Reception and... X. Summary XI. Directions for Future... References

 
For the most part, it has been assumed that the ovarian actions of estrogens are mediated via classical nuclear ERs. Specific ovarian estrogen binding (documented by way of radioligand assays) was first shown to exist in the ovaries of intact or hypophysectomized immature rats (16, 17, 18, 19, 20). Hormone binding studies by several investigators also confirmed the presence of ERs in the ovaries of intact immature hamsters, rabbits, guinea pigs, and mice (17, 21, 22). Autoradiographic studies by Stumpf (23) demonstrated the preferential localization of silver grains over nuclei of granulosa cells after in vivo administration of [³H]17ß-estradiol, thereby suggesting that the site of action of estrogen in the rat ovarian follicle is the granulosa cell. Subsequent cellular localization studies confirmed the rat granulosa cell as a site of estrogen reception (16). More recently, Arakawa et al. (20) documented (using radioligand receptor assays) the presence of ERs in antral follicles of rats.

Clemens and Richards (26) provided preliminary documentation of ER transcripts in rat granulosa cells. Specifically, primers were designed to amplify a 517-bp fragment from exon 3 to exon 5 of the rat ER. After reverse transcription of granulosa cell RNA, PCR was performed in the presence of labeled nucleotides, thereby allowing the identification and quantification of the 517-bp product and the splice variants involving exons 3–5. Subject to these limitations, no significant ER splice variants were detected. In vivo and in vitro studies revealed the granulosa cell ER transcripts to be down-regulated by 17ß-estradiol and even further reduced by the acute effects of the LH surge. As such, these findings provide a documentation of the existence of hormonally responsive ER transcripts in rat granulosa cells.

While the ER gene has been recognized since first cloned from the human breast cancer cell line MCF-7 in 1985 (27), the ERß cDNA was only isolated in 1996. The initial tissue sources of ERß transcripts were the rat prostate and ovary (28) and the human testis (29). After the cloning of ERß in rats and humans (28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46), the same was accomplished for mice (47, 48), sheep (49), cows (50), and nonhuman primates (51). Ovarian expression studies of the ERß gene have since been accomplished in rats (52, 53, 54), mice (55, 56), monkeys (57, 58), humans (59, 60), and, more recently, cows (50).

In rodent reproductive tissues, the ERß gene has been shown to be predominantly expressed in the ovary (61, 62, 63). ER transcripts, in turn, appear to predominate in the uterus, cervix, and vagina (53, 54, 64). Localization of ERß transcripts to the rat ovary was convincingly demonstrated by both in situ hybridization and RT-PCR (28, 29, 56, 65, 66, 67, 68, 69, 70, 71). Although ERß transcripts predominate in rat ovaries (68), both the ER and ERß genes have been shown to be expressed in the ovaries albeit mostly in distinct cellular compartments (54, 62, 65).

The expression of the ER gene in the mouse ovary dates back to fetal life, while ERß gene expression appears to be initiated at birth. The latter was shown to increase with age (55, 63), possibly due to the onset of follicular maturation. In the rat ovary, ERß gene expression was apparent by fetal day 14, a sharp increase being noted during the first week of neonatal life (53).

In rat ovaries, ERß proved highly localized to granulosa cells of healthy follicles, from the primary to the preovulatory stage (54, 62, 72) (Figs. 1 and 2). ERß transcripts were also expressed in theca cells although more weakly when compared with their expression in granulosa cells (65). Some studies have described scattered ERß gene expression in the ovarian stroma (65). The primordial follicle, the oocyte, and the germinal epithelium do not appear to be sites of ERß expression (73).

View larger version (174K): [in this window] [in a new window]   Figure 1. Localization of ERß mRNA in adult rat ovary during dioestrus. Bright- (a.i) and dark-field (a.ii) images show intense signal for ERß mRNA in the granulosa of ovarian follicles (F) at different stages of development, but negative signals in the corpus luteum (CL). b, Pronounced expression of ERß mRNA in granulosa cells (G) and moderate expression in theca cells (T) of healthy follicles. c, Expression of mRNA is not detectable in the granulosa cells of degenerating follicles (DF). Bar = 200 µm (a.i), 50 µm (b and c). [Reproduced with permission from Mowa and Iwanaga: J Endocrinol 165:59–66, 2000 (54 ). © the Society for Endocrinology.]

View larger version (139K): [in this window] [in a new window]   Figure 2. Localization of ERß in 23-day-old (A and B) and 60-day-old (C and D) rat ovary. Frozen sections of rat ovary were incubated with PAI P310 ERß antibody (A–C) or preadsorbed ERß antibody with the peptide immunogen (D). Note the immunostaining in follicles (F), specifically in nuclei of granulosa cells (G; panels B and C) and the lack of staining in theca cells (T), interstitial cells (I), and oocytes (O). No staining was observed when peptide-adsorbed antibody was used (panel D). Counterstained with hematoxylin; magnification, x110 (A), x800 (B), and x440 (C and D). [Reproduced with permission from Sar and Welsch: Endocrinology 140:963–971, 1999 (73 ). © The Endocrine Society.]

Some studies suggested that ERß gene expression in the rodent ovary may be under LH/hCG control (67). Indeed, preovulatory follicles, as well as newly formed corpora lutea of mice, express lower levels of ERß transcripts, a phenomenon possibly due to the preovulatory LH surge (70) (Fig. 3). The ability of LH/hCG to down-regulate ERß gene expression has also been evident in studies using cultured rat granulosa cells (67). Unlike the ERß gene, ovarian ER gene expression proved consistently uniform throughout the rat ovarian cycle. The pregnant rat ovary is a site of ER reception (54, 69).

View larger version (162K): [in this window] [in a new window]   Figure 3. Effect of hCG on protein ERß expression using immunocytochemistry. Ovary sections from rats were incubated with an ERß antiserum. ERß protein was detected in granulosa cells of small and large antral follicles in vehicle-treated (A) and PMSG-treated (B) animals. Star indicates the location of antrum in antral follicles. PMSG-treated rats were injected with an ovulatory dose of hCG and ovaries were isolated after 3 h (C), 9 h (D), 12 h (E), or 24 h (F). The expression of ERß protein in granulosa cells 9 h after hCG was reduced in large antral follicles (left, D), greatly reduced in Graafian follicles (center, D) but did not change in small antral (right, D) follicles. Similar expression was observed 12 h after hCG administration (E). One day (24 h) after hCG treatment, expression was not detected in corpora lutea (small arrowheads, F) but highly expressed in preantral (right, F) and small antral follicles. Magnification: A–F, x250. [Reproduced with permission from S. L. Fitzpatrick et al.: Endocrinology 140:2581–2591, 1999 (70 ). © The Endocrine Society.]

In summary, the ability of estrogens to influence the above mentioned activities in the rat ovary is due to the action of two receptors ( and ß) described to date. The rodent granulosa cells contain predominantly, if not exclusively, the ß-subtype, whereas the ER subtype was detected mainly in the theca layer and in the interstitium. Although ER and ERß are not expressed in the rat ovary in a complete nonoverlapping pattern, the predominance of specific subtypes in different cellular compartments within the ovary suggests that each ER subtype may be responsible for distinct downstream activities in the cell types in which they are expressed.

Although not thoroughly characterized, porcine ovarian follicles express high-affinity nuclear estrogen binding sites in classical equilibrium binding assays with in vitro dissociation constants of approximately 0.6 to 1.1 nM (74, 75). There is little or no knowledge of specific subtype ( or ß) expression and differential regulation of ERs in porcine theca or granulosa cells. However, estrogen exerts consistent actions on selected facets of follicle development and/or cytodifferentiation (see below).

	III. The Nonprimate Ovary as a Site of Estrogen Action

Top Abstract I. Introduction II. The Nonprimate Ovary... III. The Nonprimate Ovary... IV. Lessons Learned from... V. Lessons Learned from... VI. The Primate/Human Ovary... VII. Is an Estrogen-Free... VIII. The Primate/Human Ovary... IX. Estrogen Reception and... X. Summary XI. Directions for Future... References

 
A. The rat (Table 2) and mouse
In 1940, Richard I. Pencharz (76) reported on the ability of systemically administered diethylstilbestrol (DES) to produce considerable ovarian enlargement even after hypophysectomy. Moreover, the implantation of DES at the time of hypophysectomy not only maintained ovarian weight but also rendered the ovaries more responsive to exogenous gonadotropins. Indeed, a striking ovarian enlargement occurred in DES-primed animals, which were subsequently injected with CG (i.e., Antuitrin "S"). The ovaries of DES-treated animals consisted of healthy, predominantly solid (i.e., preantral), medium-sized follicles packed tightly at the expense of markedly reduced interstitial tissue. DES-primed/CG-treated animals, in turn, displayed large antral follicles, many corpora lutea, and, in two instances, hemorrhagic follicles.

View larger version (144K): [in this window] [in a new window]   Figure 4. Photomicrograph of an ovary from the immature hypophysectomized DES-treated rat. Cross-section of the entire ovary showing a near-uniform population of late secondary (type 5a and 5b) and early tertiary (type 6) follicles. The type 6 exhibits cavitation (arrows) or beginning antrum formation. Magnification, x29. [Modified from R. Sadrkhanloo et al.: Endocrinology 120:146–155, 1987 (98 ). © The Endocrine Society.]

Mice responded to DES [but less so to 17ß-estradiol cyclopenthylpropionate (ECP)] with an increase in the number of small or large preantral follicles (22), but without an increase in ovarian weight. Apart from the latter, in the few reports on the effects of estrogen in immature mice, both stimulatory and inhibitory effects on follicular development were noted (110, 111). However, since these observations were made in pituitary-intact mice, it is difficult to conclude that estrogens may have been acting directly at the level of the ovary. Fortunately, and more recently, Wang and Greenwald demonstrated unequivocally the ability of ECP, 10, 50, or 250 µg daily for 1–4 days, to stimulate the growth of preantral and antral follicles in the hypophysectomized mouse, an effect associated with a delay of follicular atresia (112). Moreover, estrogens were shown to synergize with FSH in the enhancement of follicular proliferation and differentiation as well as in the attenuation of follicular atresia (112). Taken together, it would appear that the mouse, similar to the rat, is estrogen-responsive. The precise reason(s) underlying the apparent discrepancy with earlier work remains uncertain.

B. The hamster
Although the direct effect of estrogen on rat ovarian follicular development is well established, the role of estrogen in ovarian folliculogenesis may be a species-specific phenomenon. Thus, extrapolation of the rat data to other species may not be feasible. Indeed, hypophysectomized adult hamsters, injected with estradiol benzoate, 10 µg daily for 3 days, exhibited neither increased ovarian weight nor enhanced effects of FSH on follicular development (113). Thus, hypophysectomized (or intact) hamsters may not respond to estrogen by increasing the number of large preantral follicles as reported in the rat. Similar results were obtained for the intact or hypophysectomized immature hamster (17). It is possible that for the immature and adult hamster, unlike the rat, estrogens do not play a major role in the recruitment of large ovarian preantral follicles. In this same context, Hutz et al. (21) set out to observe the response of granulosa cells from gonadotropin-primed hamsters to treatment with DES or 17ß-estradiol under in vitro circumstances. Interestingly the application of DES inhibited the accumulation of estrogen regardless of the presence or absence of FSH in the culture medium. In contrast, the combination of 17ß-estradiol plus FSH augmented the accumulation of progesterone, which clearly argues that estrogen is facilitating granulosa cell differentiation. These findings were interpreted to mean that estrogens can be nonstimulatory or inhibitory to the function of hamster granulosa cells in vitro in parallel to effects shown in vivo.

Since the role of estrogens in hamster ovarian folliculogenesis is uncertain, comparisons between hamster and other animal models as well as conclusions on the effect of estrogen on ovarian follicular development are hindered by the paucity of available data.

C. The rabbit and guinea pig
Immature rabbits treated with ECP or DES for 3 days failed to increase ovarian weight but did increase the number of small or large preantral follicles (22). In contrast, estrogen-treated guinea pigs displayed a significant increase in ovarian weight, a phenomenon attributable to an increase in the number of large antral follicles (22).

D. The pig
Experimental studies support a thesis of multipotential actions of estrogen on follicular growth and cytodifferentiation in the pig (74, 114, 115, 116, 117, 118, 119). However, since much of this knowledge is based on descriptive in vitro studies, which may involve risk, interpretative complications due to 1) the presence of serum (120, 121), 2) failure to distinguish atretic from healthy follicles, 3) assessment of a limited degree of estrogen’s interactions with other regulators of follicle function, and (4) focus on growth and cytodifferentiative features of the granulosa, but not theca cell, it is difficult to make definitive assertions on the role of estrogen in the swine ovary.

Whereas in vivo estrogen administration in the (hypophysectomized) rat clearly supports follicular and ovarian growth, this inference is confounded in the pig by well documented granulosa cell-proliferative responses at least in vitro to insulin-like growth factor I (IGF-I), the production of which by granulosa cells can be driven by GH alone or estrogen and FSH combined (122, 123, 124, 125). Moreover, IGF-I or estrogen each is capable (alone or in concert with FSH) of augmenting indices of in vitro granulosa cell proliferation in the pig, e.g., inducing proliferating-cell nuclear antigen (PCNA) expression and increasing tritiated thymidine incorporation or cellular DNA content (126, 127).

The full physiological in vivo role(s) of estrogen in modulating ovarian cellular growth will be challenging to unmask in swine, since multiple intrafollicular factors also appear to control granulosa cell growth, and their possible interactions with estrogen have not yet been explored, e.g., inhibitory [tumor necrosis factor- (TNF) (126)], or stimulatory [relaxin (128), transforming growth factor-ß (TGFß) (129), endothelin-1 (129), epidermal growth factor (EGF) (127), cAMP analogs (127), and T₄ (130)].

Nonetheless, because the majority of correlative studies document an inverse relationship between intrafollicular estradiol concentrations and one or more cytological measures of atresia (e.g., Ref. 131), several important considerations arise, namely whether 1) estrogen is obligatory to maintain healthy late-follicle development; or, conversely, 2) failure of follicular development, predicated on whatever mechanistic grounds, is heralded by loss of estrogen-synthesizing capacity. Both views are supported, but not proven, by the concomitant waning of estrogen synthesis and other biosynthetic functions in early follicular atresia in the pig, e.g., inhibin production or FSH-stimulated cAMP accumulation (131). Moreover, aromatase activity in the pig is controlled by multiple factors other than FSH acting alone [e.g., inhibitory: cytokines, such as interleukin-6 (IL-6) (132), PRL (133), extracellular purines (134), insulin-like growth factor binding protein-3 (IGFBP-3) (135), and high concentrations of LH (136); or stimulatory: IGF-I or EGF (137)].

In contrast to the sparse evidence for follicle (or granulosa cell) growth-promoting actions of estrogen in the pig (above), significant trophic actions of estrogen on multiple endpoints of granulosa cell cytodifferentiation are well recognized. For example, estradiol acts trophically in vitro as a potent biological amplifier of PRL, FSH, and IGF-I or IGF-II’s stimulation of progesterone production by (immature) porcine granulosa cells in monolayer sparsely serum-supplemented or serum-free first-passage cultures (74, 118, 119, 138, 139, 140, 141, 142, 143, 144, 145, 146, 147, 148, 149, 150, 151). The mechanisms underlying such singular and synergistic actions of estradiol include enhanced low density-lipoprotein (LDL)-receptor-expression, increased cholesterol uptake and utilization (152), augmented cytochrome P450_scc enzymic activity (148, 150, 152), increased conversion of pregnenolone to progesterone, heightened de novo synthesis of cellular cholesterol via 3-hydroxy-3-methyl-glutaryl-coenzyme A reductase (146), and augmented steroidogenic acute regulatory protein expression, without evident inhibition of progesterone’s metabolism to 5-dihydroprogesterone (74). These in vitro trophic actions of estradiol occur over 36–48 h or longer without any measurable proliferative effects. Rather, inhibition of porcine granulosa cell proliferation with a noncompetitive (suicide) inhibitor of polyamine biosynthesis amplifies the daily rate of in vitro progesterone-expressed biosynthesis per granulosa cell (153), suggesting an inverse relationship between granulosa cell proliferation and cytodifferentiation (154).

Estrogen’s modulation of progesterone biosynthesis by pig granulosa cells is bipotential in a time-dependent sense (74, 75, 118, 119, 138, 139, 140). Short-term (2–18 h) exposure to estradiol consistently inhibits progesterone biosynthesis, apparently by directly antagonizing pregnenolone’s conversion to progesterone via 3ß-hydroxysteroid dehydrogenase (74, 153). This direct action is mimicked by a nonmetabolizable estrogen, moxestrol, and is not antagonized by the antiestrogen, keoxifene (144). Since inhibition can be reproduced in broken cell (microsomal) preparations, it is likely that direct steric inhibition of enzyme catalysis occurs. The mechanisms underlying delayed (34–36 h) escape from this acute inhibition are not yet known (74, 118, 119, 138, 139). However, an analogously rapid inhibition of steroidogenesis by estrogen is described in pig theca cells (155, 156, 157). Such steroidogenic autoinhibitory actions of estrogen may be relevant as the preovulatory LH surge unfolds, when intrafollicular 17ß-estradiol concentrations approach or exceed 1 µg/ml (74, 158), thereby (inferentially) limiting a premature elevation of progesterone concentrations before corpus luteum formation and/or thereby restraining the preovulatory (theca cell-derived) androgen surge.

E. Possible interactions of estrogen with other putative ovarian regulators
Whereas estrogen is the primary focus above, its critical interactions with other known and putative intrafollicular regulators will be essential to unravel the complete understanding of the ovarian follicular development in the pig and other species (130, 159, 160). For example, a reciprocal relationship emerges between growth (cell proliferation) and cytodifferentiation in response to certain modulators, e.g., endothelin-1 (129, 161, 162), polyamines (153), and TNF (126). This observation, while not universal (IGF-I tends to promote both granulosa cell proliferation and differentiation) illustrates the potential complexity inherent in in vivo interpretations of estrogen’s acting in combination with other potent regulators of folliculogenesis, such as the intraovarian IGF-I system (145). Estrogen-growth factor interactions are of particular interest, when folliculogenesis is viewed as a continuum of changing cellular commitment to replication vs. cytodifferentiation.

How estrogen interacts, if at all, with specific cell-surface adhesive proteins, such as connexin, laminin, or integrin, in their presumptive maintenance of commitment to granulosa cell lineage (120, 122, 163) is likely of significance in the maturing follicle, as well as possibly in the earliest stages of preantral follicle formation (164, 165).

Estrogen’s presumptive modulation of the calcium-calmodulin, C-kinase, and A-kinase effector pathways in granulosa and theca cells will also require further study (125, 140, 141, 166, 167, 168, 169, 170, 171, 172, 173, 174, 175, 176, 177, 178, 179, 180, 181, 182, 183, 184). In this light, it will be critical to establish whether, when, and how estrogen participates in rescuing follicles from early or incipient atresia (131, 185). Differential hybridization of pig granulosa and theca cell mRNAs collected in untreated vs. estrogen-treated conditions, as accomplished for FSH recently (186), could be one step toward addressing this difficult issue. In vitro whole-follicle cultures also may be useful in this context (187).

Given the pivotal role inferred to date for the IGF-I and IGF-II (and their associated binding proteins) systems in follicle development, exactly how estrogen impacts the multifaceted actions of IGF-I and IGF-II in the pig will require further clarification (135, 149, 150, 151, 152, 181, 182, 188, 189, 190, 191). In addition, estradiol’s interactions with FSH receptor-mediated cytodifferentiative actions on granulosa cells should be evaluated further, given the unequivocally central role of this gonadotropin in the follicle development in this and other species (190).

Further knowledge of how estradiol modulates intrafollicular production and actions of the multiple cytokines and activin/inhibin glycoproteins (and the structurally unrelated follistatin) may also help better elucidate estrogen’s role in folliculogenesis in the pig. Although of interest for other investigative purposes, the recent development of pig granulosa cell clonal lines is not likely to be so rewarding in dissecting the foregoing broader physiological issues (192). Lastly, we suggest that the potent luteotropic effects of estrogen in the pig, albeit well supported by available data in this species, require better mechanistic understanding (193, 194, 195, 196, 197). The foregoing queries will eventually also need to be addressed with respect to the nature and mechanisms of estrogen’s modulation of theca and/or interstitial cell function in the ovary. The latter proposition assumes our thesis that folliculogenesis progresses under the dual orchestration of the intrafollicular and perifollicular milieus, which likely jointly coordinate follicle development.

	IV. Lessons Learned from the Estrogen-Resistant Model-Estrogen Receptor Null Mutants (ERKOs)

Top Abstract I. Introduction II. The Nonprimate Ovary... III. The Nonprimate Ovary... IV. Lessons Learned from... V. Lessons Learned from... VI. The Primate/Human Ovary... VII. Is an Estrogen-Free... VIII. The Primate/Human Ovary... IX. Estrogen Reception and... X. Summary XI. Directions for Future... References

 
The first and only known case of clinical estrogen insensitivity in man was reported by Smith et al. (198). This male patient displayed no detectable response to estrogen administration due to an inactivating mutation of the ER gene. The patient was tall and normally masculinized. However, estrogens proved to be of import for bone maturation and for mineralization. Although no examples of estrogen resistance exist in the primate/human female, the recent flurry of activity in this area might indicate that such an occurrence is not an impossibility. Clearly, once such an individual is identified, significant lessons could be learned with respect to the relative importance of estrogens to ovarian follicular maturation in the human female.

Short of such primate/human models, note must be made of the recent works reporting the generation null mutants for the ER (199, 200, 201, 202, 203, 204, 205) and ERß (206) genes by way of gene targeting (homologous recombination) technologies (207). Compound null mutants were also generated. A detailed account of this work was recently offered (207). Therefore, only a brief discussion of the relevant points will appear here.

ER-/- mice survived to adulthood featuring a normal gross phenotype. However, null-mutant female mice proved infertile, displaying hypoplastic uteri and hyperemic ovaries with no detectable corpora lutea (205). Histological sections of ovaries from homozygous null mutants revealed no defects in germ cell formation or migration. Further, developmental progression through the primordial, primary, and antral follicle stages appeared normal. However, functional maturation to preovulatory follicles was arrested resulting in atresia or in "anovulatory follicles," which in many cases formed large, hemorrhagic cysts containing few, if any, granulosa cells. Stated differently, ovarian follicles of adult ER-/- mice may progress to a "Graafian" state albeit in the face of abnormal stratification of granulosa cells, some areas of the follicle being surrounded by multiple layers of cells, other regions featuring a single layer of squamous-appearing cells (205, 208) (Fig. 5). Estrogen actions, such as the attenuation of apoptosis or the amplification of the LH receptor content in granulosa cells of antral follicles, appear preserved, probably due to mediation by the ERß receptor.

View larger version (56K): [in this window] [in a new window]   Figure 5. Histological examination of adult wild-type (WT) and ERKO ovaries. A, Histological examination of normal adult WT ovaries revealed many corpora lutea (CL) (stars). Magnification bar = 500 µm. B, Ovaries from ERKO female mice contained many hemorrhagic ovarian cysts (asterisks). However, occasional Graafian ovarian follicles (arrow) were present. Magnification bar = 500 µm. C, Higher magnification of Graafian ovarian follicle from adult ERKO ovary depicted in panel B reveals that there was differentiation of granulosa cells with cumulus cells surrounding the ovary and follicular antral fluid in the central portion of the ovarian follicle. However, there was abnormal stratification of granulosa cells, with a portion of the follicle lined by a single layer of cells (bracketed area). Magnification bar = 200 µm. [Reproduced with permission from C. S. Rosenfeld et al.: Biol Reprod 62: 599–605, 2000 (208 ). © Society for the Study of Reproduction.]

However, the role of LH in the evolution of the above phenotype can be questioned, given two other null mutant models in which the hemorrhagic cyst phenotype is absent (in the face of chronically elevated LH levels). For example, the null mutant for the FSHß-subunit gene (215), otherwise replete with LH, proves the importance of FSH in follicular cyst formation. Similarly, the null mutant for the P450- aromatase gene (216), characterized by high circulating levels of both LH and FSH, but absent estradiol, points out a role for estradiol in the genesis of the (hemorrhagic) cyst phenotype.

To test whether ER is required for ovulation and corpus luteum formation, gonadotropins were used to superovulate immature ER-/- mice and wild-type siblings (208, 214). Gonadotropin therapy resulted in ovulation in both ER-/- and wild-type mice. However, fewer gonadotropin-treated null mutants ovulated. In addition, ER-/- mice yielded significantly less oocytes. Surprisingly, ovulated/ruptured ovarian follicles of null mutants developed into corpora lutea of apparent normal morphology boasting 3-fold the concentrations of serum progesterone as compared with controls. However, adult ER-/- mice could not be induced to ovulate, probably due to the elevated circulating levels of LH and the development of hemorrhagic ovarian cysts (200).

Null mutants for the ERß gene survive to adulthood and exhibit a phenotype distinct from that of their ER-/- counterparts. Unlike ER-/- mice, the ERß-/- female mouse proved fertile, albeit with apparently less frequent ovulation (206). Consequently, fewer corpora lutea were apparent, the litter size being smaller than in wild-type mice. Not unlike the ER-/- mice, gonadotropin-treated ERß-/- mice responded with fewer oocytes released than their wild-type counterparts (206).

Histological analysis of ovaries of ERß-/- mice was mostly normal, except for an increased number of early atretic follicles and the sparse presence of corpora lutea, suggesting arrested folliculogenesis. Superovulation disclosed a reduced ovulatory capacity of ERß-/- mice as compared with their wild-type counterparts. The histology of the ovaries from superovulated ERß-/- mice revealed the presence of numerous unruptured preovulatory follicles, indicating deficiency in the response to the gonadotropin surge (hCG). The relative deficit of spontaneous ovulation in the ERß-/- mice may be due to diminished up-regulation of ovarian progesterone receptor (PR) levels by gonadotropins, or an alteration in gonadotropin synthesis or secretion (207, 217). Dupont et al. (218) described a similar phenotype for ERß null mutant, but their data reveal that only half of the superovulated female ERß null mutants ovulate, in contrast to the study of Krege et al. (206) who reported that 80% of the superovulated animals ovulated.

Recently, mice lacking both ER and ERß were generated (219). ERß-/- mice exhibited normal reproductive tract development but proved infertile. Ovaries of adult ERß-/- mice exhibited follicle transdifferentiation to structures resembling seminiferous tubules of the testis, including Sertoli-like cells expressing Mullerian inhibiting substance, sulfated glycopropein-2 and Sox9 (biochemical markers of Sertoli cell differentiation). In some follicles, a recognizable but degenerating oocyte was present, whereas others featured no evidence of germ cells. These findings indicate that both receptors are required for the maintenance of germ and somatic cells in the postnatal ovary. Interestingly, serum LH levels were higher than those observed in ER-/- mice, suggesting that both ERs are required for estradiol-mediated regulation of LH secretion (219).

In summary, these findings suggest that prenatal reproductive tract development in the mouse can occur in the absence of ER or ERß, an observation consistent with traditional theory. However, reproductive function is undoubtedly affected. Since, ovarian folliculogenesis, ovulation, and corpus luteum formation can occur in the absence of ER or ERß, albeit less effectively than in wild-type mice, it might be suggested that neither one of the known ERs is essential, but that each may play a facilitatory role in ovarian follicular development and maturation.

	V. Lessons Learned from the Targeted Disruption of the CYP-19 (Aromatase) Gene-Aromatase Null Mutant (ArKO)

Top Abstract I. Introduction II. The Nonprimate Ovary... III. The Nonprimate Ovary... IV. Lessons Learned from... V. Lessons Learned from... VI. The Primate/Human Ovary... VII. Is an Estrogen-Free... VIII. The Primate/Human Ovary... IX. Estrogen Reception and... X. Summary XI. Directions for Future... References

 
The formation of estrogens from C₁₉ steroids is catalyzed by aromatase cytochrome P450, the product of Cyp 19 gene. Null mutant mice for aromatase gene (ArKO) were generated (216), thereby affording the opportunity to examine the role of estrogen in the follicular development in the mouse ovary. Upon reaching sexual maturity, female ArKO mice, born phenotypically normal, develop a male body habitus with excessive internal fat deposition, underdeveloped uteri, and ovaries lacking corpora lutea. The clitoral glands, mammary glands, and gonadal fat pads enlarge. Histological evaluation of the ovaries reveal the presence of many large follicles filled with granulosa cells and evidence of antrum formation, but no corpora lutea. As expected, testosterone and LH levels are markedly elevated. The high testosterone levels presumably reflect stimulation of the theca interna cells by LH (216, 220, 221, 222, 223, 224, 225, 226, 227, 228, 229, 230). Although estrogen was not a prerequisite for the reinitiation of follicle growth (from the point of primordial follicle up to the antral follicle stage), note was made of a block of follicular development and absent corpora lutea. The ovarian phenotype degenerated with age upon the appearance of hemorrhagic cystic follicles and the loss of secondary and antral follicles coincident with the infiltration of macrophages and with stromal hyperplasia (228, 231). In summary, the ArKO female mouse is infertile, as a consequence of disrupted folliculogenesis and failure to ovulate.

	VI. The Primate/Human Ovary as a Site of Estrogen Reception

Top Abstract I. Introduction II. The Nonprimate Ovary... III. The Nonprimate Ovary... IV. Lessons Learned from... V. Lessons Learned from... VI. The Primate/Human Ovary... VII. Is an Estrogen-Free... VIII. The Primate/Human Ovary... IX. Estrogen Reception and... X. Summary XI. Directions for Future... References

 
Essential to any discussion of a local paracrine/autocrine role for estrogens in primate/human ovarian physiology must be a review of the evidence relevant to the possibility that the primate/human ovary may be a site of estrogen reception.

A. Molecular probing (Table 3)
Billiar et al. (232), relying on Northern blot analysis, made use of a cDNA probe corresponding to the 1.8-kb open reading frame of the ER expressed by the MCF-7 human breast cancer cell line (232). A 7-kb ER mRNA species was detected when using a poly(A)⁺-enriched mRNA fraction representing whole ovarian material from two animals. Unfortunately, the use of whole ovarian material all but precluded the localization of ER mRNA to discrete follicular or perifollicular components. Consequently, one cannot rule out that the positive signal reflects germinal epithelial contamination.

More recent reports by Hurst et al. (234, 235), using RT-PCR technology applied to highly luteinized human "granulosa cell" mRNA, provide data in apparent conflict with those reported by Wu et al. (233). To identify the possibility of genomic DNA contamination, the oligonucleotide primers were so designed as to correspond to base pairs 570–852 in the B and C domains of the human ER cDNA, a stretch of nucleotides known to span intron 1. Southern blotting of the amplified products with a ³²P-labeled ER mRNA probe confirmed the existence of ER transcripts (Fig. 6). The data suggest the existence of ER transcripts in follicular aspirates consisting largely of luteinized granulosa cells. However, since the granulosa cells studied were obtained via transvaginal needle aspiration of ovarian follicles in the course of oocyte retrieval in conjunction with in vitro fertilization procedures, the signal detected may correspond to cells other than granulosa cells known to be present in follicular aspirates (e.g., representatives of the white blood cells series). Consequently, there exists the strong possibility of contamination by vaginal tissues or by ovarian germinal epithelium, which may be ER-positive, thereby reducing the level of confidence in the conclusions from this study.

View larger version (70K): [in this window] [in a new window]   Figure 6. Agarose gel demonstrating the amplification of the 282-bp ER fragment in human granulosa cells (GC), but not in the negative control human glioblastoma cell line, JHN J889H. The ß-actin (b-A) fragment is amplified from both cell types. [Reproduced with permission from B. S. Hurst et al. : J Clin Endocrinol Metab 80:229–232, 1995 (234 ). © The Endocrine Society.]

Revelli et al. (237) also undertook to evaluate the expression of ER transcripts in the ovaries of 25 healthy eumenorrheic women. Ovarian biopsies were taken in different phases of the menstrual cycle during laparotomy or operative laparoscopy performed for extraovarian benign diseases. Using in situ hybridization targeted at ER transcripts, Revelli and associates documented ER transcripts in 17.1% of primordial follicles. The proportions of preantral and antral follicles positive for ER transcripts were 30.7 and 37.5%, respectively. No more than 25% of theca cells proved ER transcript-positive. Active corpora lutea stained positive for ER transcripts in 50% of the cases. Corpora albicantes always stained negative. In all subjects, the stroma surrounding both follicles and corpora lutea contained several fibroblast-like cells that stained positive for ER transcripts. Oocytes and blood vessels stained negative in all cases.

Brandenberger et al. (238) compared the expression profiles of ER and of ERß transcripts in the midgestational human fetus by semiquantitative RT-PCR. ER was detected in the human fetal ovary. However, the relative amounts of ER transcripts in the human fetal ovary were substantially reduced as compared with the levels observed in the uterus. In contrast, significant amounts of ERß transcripts were present in fetal ovaries. In relative terms, ERß transcripts were far more prominent than their ER counterparts. The presence of ERß transcripts in ovarian granulosa cells of adult women was also convincingly demonstrated (239). Similar amounts of ER and ERß mRNA were reported for normal ovaries in all age groups from 33 to 75 yr (59). Luteinized granulosa cells also expressed a significant level of ERß mRNA (59).

Note is also made of the contribution of Pau et al. (240), who were able to observe (by means of RT-PCR and in situ hybridization) weak to moderate signals of ERß and strong signals of ER mRNA within the granulosa layer of primate ovarian follicles. Similarly, whole ovarian mRNA yielded appropriate amplicons corresponding to ER and ERß after RT-PCR.

Pelletier et al. (57), in an effort to clarify the expression of ERß in the reproductive organs of primates, performed in situ hybridization studies in ovaries of adult female Cynomolgus monkeys (Macaca fascicularis). ERß mRNA localized to granulosa cells of follicles at different stages of development, including small growing and secondary (antral) follicles. Primordial follicles, however, were devoid of ERß expression. The theca interna cells were also strongly labeled. It was not possible to identify any preovulatory follicle. The corpora lutea appeared to be a site of ERß gene expression, whereas interstitial cells were consistently negative. A strong autoradiographic reaction obtained for the ovarian capsule, i.e., the surface epithelium and the stroma cells.

Mention must also be made of studies on the regulation of ER and ERß transcripts in the superovulated primate follicle (241). ER transcripts did not change whereas ERß transcripts decreased 12–36 h after the ovulatory stimulus. Steroid ablation reduced ER transcripts 12 h after hCG, an effect partially reversible by progestin replacement. ERß transcripts were unaffected by steroids. These data demonstrate hCG-initiated, steroid-dependent (ER) and -independent (ERß) expression of receptor transcripts in primate granulosa cells during the periovulatory interval. Differences in patterns of expression may relate to the possible diverse roles of steroid hormones in periovulatory events.

Recent in vitro studies on ER and ERß gene expression in human granulosa-lutein cells have improved the understanding of the hormonal regulation of ERs in the ovary (242). Although transcripts and protein are identified for both ER and ERß in human granulosa cell cultures, ER is expressed at a lower level than ERß. Basal expression studies indicated that ER mRNA levels remained unchanged, whereas ERß mRNA levels increased with time in culture. Treatment with hCG significantly attenuated the expression of both receptors. Similarly, treatment with GnRH led to inhibition of the ER and ERß transcripts and proteins. These results suggest that ERß more than ER may play a dynamic role in regulating corpus luteum formation in the human.

It has been observed that both human ER genes give rise to a significant number of mRNA isoforms, which, in turn, exhibit differential expression among the tissues in which they reside (41, 243). In human ovaries, the C and F mRNA isoforms of the ER gene appear to be the major forms detected among six human mRNA isoforms, A–F. Moore et al. (41), in evaluating the expression of the five mRNA isoforms for the human (h) ERß gene (hERß 1–5), was able to identify hERß-1 and -2 as most abundant in the normal ovary, whereas hERß-4 could not be detected (41). In addition, a novel human estrogen receptor ß isoform, ERßcx, has been identified: one that is truncated at the C-terminal region but possesses an extra 26 amino acids due to alternative splicing. The ERßcx transcript is expressed in the adult human ovary (244). Interestingly, ERßcx displays no ligand binding ability and fails to form any shifted complex in gel shift assays. Moreover, in a transient expression assay, ERßcx shows no ligand-dependent transactivation ability of a basal promoter and also cannot interact with a cofactor, TIF1, in the presence or absence of 17ß-estradiol. ERßcx preferentially forms a heterodimer with ER rather than with ERß, inhibiting DNA binding by ER. Thus, ERßcx may potentially inhibit ER-mediated estrogen action.

In summary, the nonhuman primate ovary is a site of expression of ER transcripts, mainly at the level of the germinal epithelium (236). The granulosa cell layer may also be a site of ER gene expression (240). The adult human ovary, in turn, may express ER transcripts at the level of the oocyte (233) and the luteinized granulosa cell (234) and, as recently documented, the germinal epithelium (60). Revelli et al. (237), however, failed to document ER transcripts at the level of the human oocyte; instead, they reported ER transcript positivity in a portion of follicular granulosa cells and in a minority of theca cells. The status of the oocyte notwithstanding, it would appear that the adult human ovary is a site of ER expression. ER transcript positivity was also noted in 50% of active corpora lutea (237). The fetal human ovary is a site of ER and ERß gene expression wherein the ERß transcripts are more abundant (238). Recent probing establishes the presence of ERß transcripts in the adult human ovary at the level of the granulosa cell (59, 239) and the germinal epithelium (60). In a recent study, both ER and ERß transcripts colocalized to the human corpus luteum (245). In the nonprimate ovary, ERß transcripts are consistently expressed in granulosa cells of growing follicles as well as in the theca cell layer and germinal epithelium (57). Given the recent discovery of the inhibitory version of an ER, i.e., ERßcx (244), it is reasonable to presume that the net estrogenic impact at the level of the human ovary may be determined by the relative expression and interaction of the various ER subtypes.

B. Immunohistochemical studies (Table 4)
The first reported attempt to identify ERs in the primate ovary was performed by Hild-Petito et al. (246) using adult rhesus (Macaca mulatta) and cynomolgus (Macaca fascicularis) monkeys. Ovaries were collected from adult animals in the early, mid, and late follicular phase as well as in the luteal phase of the menstrual cycle (n = 3–6 ovaries per stage). Specific monoclonal antibodies, directed against the human ER (H222 and D75), were employed (247, 248). Surprisingly, immunoreactive ER was only localized to the ovarian germinal epithelium, an observation applicable to all stages of the cycle (Fig. 7). Indeed, all the other structures in the ovary were immunonegative, regardless of the phase of menstrual cycle. In contrast, progesterone receptors (PRs) were readily demonstrated using similar techniques in many of the (ER-negative) ovarian tissues studied. As such, these findings lend further credence to the experimental conclusions by arguing against the possibility that the apparent paucity of ER reflects technical shortcomings.

View larger version (82K): [in this window] [in a new window]   Figure 7. Frozen sections of ovarian cortex from one monkey during the early follicular phase of the menstrual cycle. Germinal epithelium (e) exhibited intense nuclear staining when the antibodies against ER(A) or PR(B) were used. When the first antibody was replaced with antibody to antigen B of timothy grass pollen (C), the nuclear reaction product was absent. Sections were counterstained with hematoxylin. Magnification, x320. Bar is 31 µm. [Reproduced with permission from S. Hild-Petito et al.: Endocrinology 123:2896–2905, 1988 (246 ). © The Endocrine Society.]

A similar immun