Ovarian Steroids: The Good, the Bad, and the Signals that Raise Them

Abstract

Ovarian steroid production and subsequent local steroid-mediated signaling are critical for normal ovarian processes, including follicle growth, oocyte maturation, and ovulation. In contrast, elevated steroidogenesis and/or increased steroid signaling in the ovary can lead to profound ovarian pathology, such as polycystic ovarian syndrome, the leading cause of infertility in reproductive age women. Through the use of several in vitro and animal models, great strides have been made toward characterizing the mechanisms regulating local steroid production and action in the ovary. Examples of this progress include insights into luteinizing hormone (LH)- and growth factor-mediated signaling, steroidogenic acute regulatory protein (StAR) activation, and both genomic and nongenomic steroid-mediated signaling in somatic and germ cells, respectively. The following review will address these advances, focusing on how this rapidly expanding knowledge base can be used to better understand female reproduction, and to further improve treatments for common diseases of infertility.

Introduction

Ovarian steroid production is critical for the normal development and function of several distant tissues in women, including the uterus, breast, skeleton, and brain. Importantly, steroids produced from the ovary also have local effects that are essential for normal ovarian physiology. For example, estradiol-mediated signaling through the estrogen receptors ERα and ERβ is critical for follicle development ^1–3, as mice with reduced expression of one or both or these receptors have abnormal follicle growth. Likewise, androgen signaling via the androgen receptor (AR) also appears to be important for normal follicle development, as mice lacking the AR have reduced follicle growth and premature ovarian failure ^4–6. Finally, local progesterone production and actions at the time of the luteinizing hormone (LH) surge are crucial for ovulation, as mice lacking progesterone receptors (PRs) do not ovulate in response to gonadotropin⁷.

While normal ovarian steroid production is critical for the maintenance of standard ovarian development and function, excess ovarian steroid secretion and subsequent local steroid signaling can lead to significant ovarian pathology. The most common clinical example of increased steroid signaling in the ovary leading to abnormal function is polycystic ovarian syndrome (PCOS), which is characterized by excess ovarian (and sometimes adrenal) androgen production, abnormal follicle growth, and infertility due in large part to anovulation⁸. The following review will examine the intracellular signals that regulate ovarian steroid production, as well as the physiologic and pathologic consequences of intra-ovarian actions modulated by these steroid hormones.

Normal steroidogenesis in the ovary

The Two Cell Model

Normal ovarian steroid production is a tightly regulated and complex process that involves several different signaling pathways in multiple kinds of cells (Fig. 1). A typical mammalian ovarian follicle consists of at least four cell types. Each follicle contains an oocyte surrounded by cumulus granulosa cells. These cells are then bounded by outer mural granulosa cells, which in turn are surrounded by theca cells ^{9, 10}. Steroidogenesis in most mammals (including humans and mice) appears to occur via the two cell/two gonadotropin model whereby androgens are synthesized from cholesterol in LH-stimulated theca cells, followed by conversion to estrogens in follicle stimulating hormone (FSH)-exposed granulosa cells. LH receptors and the enzyme CYP17, which converts pregnenolone and progesterone to dehydroepiandrosterone (DHEA) and androstenedione, respectively, are expressed primarily in theca cells, while FSH receptors and aromatase (CYP19), which converts androgens to estrogens, are expressed mainly in granulosa cells. Notably, LH receptors are not exclusively found in theca cells, as their expression levels in mural granulosa cells have been noted to increase in response to FSH just prior to the LH surge ¹¹.

Figure 1.

Gonadotropins trigger multiple signals that regulate steroidogenesis. LH-induced activation of the LH receptor on theca, and likely mural granulosa, cells triggers several signaling pathways that can regulate steroid production. Gα_s-mediated increases in cAMP lead to elevated PKA activity, which in turn can increase StAR and CYP17 expression and activity via mechanisms that are still not well understood. Concomitant Gβγ-mediated signaling can activate the Src/Ras/MAPK signaling cascade, which in turn has been shown to both increase and decrease StAR and CYP17 activity, depending upon the cell culture system. Of note, LH-induced Gβγ and possibly Gq signaling can also trigger PLC activation, which may affect steroidogenesis (not shown). Finally, through yet unknown mechanisms, LH receptor signaling can trigger activation of MMPs, which in turn may cleave membrane-EGF moieties to release soluble EGFs (possibly amphiregulin, epiregulin, and/or HB-EGF). These soluble EGFs can then activate the EGF receptor on granulosa, and possibly theca, cells to increase StAR activity via mechanisms that have yet to be characterized. Finally, similar to the LH receptor, activation of the FSH receptor on granulosa cells increases intracellular cAMP and activates PKA, which in turn can increase StAR and CYP19 (aromatase) expression and activity.

Interestingly, both theca and granulosa cells express steroidogenic acute regulatory protein (StAR, see below), CYP11A1 (side chain cleavage enzyme), and 3β-HSD. Both of these cell types are therefore capable of making pregnenolone and/or progesterone from cholesterol. However, in the follicular phase, the relatively avascularized granulosa cells see limited oxygen or cholesterol, thus low amounts of these steroids are produced. In contrast, after exposure to gonadotropins, the granulosa cells become “luteinized,” and are then able to synthesize large amounts of pregnenolone and progesterone from cholesterol ^{9, 10}. Notably, since these granulosa cells lack CYP17, they cannot metabolize progestins to androgens, as would occur in CYP17-expressing theca cells. Thus, most ovarian progesterone production likely comes from granulosa rather than theca cells. Since progesterone is necessary for normal ovulation ¹², and appears to be capable of promoting oocyte maturation ¹³, understanding the mechanisms by which LH-induced signaling in theca cells leads to progesterone production in granulosa cells is essential.

StAR Protein

In order to understand how progesterone production is regulated in the ovary, an appreciation of how StAR protein is expressed and activated is necessary. StAR is the primary transporter of cholesterol from the outer to inner mitochondrial membrane, where most of the steroidogenic enzymes are located. Thus, StAR is a critical, rate-limiting regulator of steroid production in all major steroidogenic tissues. For example, StAR is essential for corticosteroid production in the adrenal gland, as humans or mice lacking StAR activity are born with adrenal insufficiency requiring corticosteroid replacement ^{14, 15}. Similarly, StAR is critical for normal androgen production in the testes, as male humans and mice lacking active StAR have external female genitalia due to insufficient androgen production in utero. However, males still produce enough testosterone to retain their internal Wolffian remnants, suggesting that small amounts of StAR-independent steroidogenesis may still occur in the testes during embryogenesis and perhaps even shortly after birth. Finally, StAR is needed for normal ovarian steroid production, as adult female mice and humans lacking functional StAR have ovarian failure. Similar to the testes in embryonic males lacking StAR, ovarian function is still partially intact in young females; however, unlike the males, females lacking StAR can continue to make sex steroids throughout the pre-pubertal period, with hormone levels usually declining shortly after the onset of puberty ¹⁴. In fact, some human females lacking StAR activity still progress through puberty, and even briefly cycle, before amenorrhea ensues ¹⁵.

A “two hit” model has been described to explain the variable steroidogenic organ failure in animals and people lacking functional StAR ¹⁵. This model proposes that low-level StAR-independent steroidogenesis initially occurs in cells from all steroidogenic tissue. With time, however, these cells accumulate cholesterol that cannot be metabolized, resulting in eventual glandular destruction. With regards to the adrenals, this “second hit” primarily occurs during fetal development, as fetal adrenal tissue normally makes steroids; thus, adrenal cells lacking StAR accumulate significant lipids during embryogenesis. Similarly, the testes make steroids both in utero and just after birth; thus, Leydig cell damage occurs early in males. In contrast, the ovary remains relatively dormant until puberty; thus this gland is often preserved until that time. Once steroidogenesis begins, however, ovaries lacking StAR rapidly accumulate cholesterol and are destroyed like the adrenal gland and testes. Whether early StAR-independent steroid production in the ovary is sufficient to regulate steroid-mediated processes such as follicular growth and oocyte maturation has yet to be determined.

LH-induced signaling molecules that may regulate ovarian steroidogenesis

StAR protein expression and activity is regulated by many factors in steroidogenic tissue, starting with activation of the LH receptor in theca cells and possibly mural granulosa cells. LH-triggered signaling via its G protein-coupled receptor induces several important second messengers that have been implicated as regulators of StAR activity and/or steroidogenesis in general, including the following (Fig. 1):

cAMP and PKA

While many intracellular signaling pathways have been linked to LH-induced ovarian steroid production; gonadotropin-induced elevation of cAMP is one of the few such pathways that has been definitively proven to be important. Both the FSH and LH receptors couple to Gα_s when activated by their respective ligands. Gα_s in turn stimulates adenylyl cyclase, leading to an elevation in intracellular cAMP levels and subsequent activation of protein kinase A (PKA) ⁹. Elevation of intracellular cAMP in theca or granulosa cells by gonadotropins, forskolin, or membrane-permeable dibutryl cAMP (dbcAMP) promotes steroidogenesis by increasing expression of StAR as well as by increasing StAR activity via phosphorylation on serine 195 ^{9, 16}. Furthermore, cAMP regulates CYP19 (aromatase) and CYP17 expression and activity in granulosa and theca cells, respectively, which contributes to estrogen and androgen production ¹⁰.

Phospholipase C (PLC)

Although primarily coupling with Gα_s, evidence suggests that the LH and FSH receptors can also activate Gq and Gα_i in response to ligand ¹⁷. Both Gq and Gβγ are both known to activate PLC, which then triggers the production of second messengers inositol-1,4,5-triphosphate (IP₃) and diacylglycerol (DAG). These molecules in turn increase intracellular calcium and activate PKC, respectively. Under some conditions, PKC and calcium may inhibit adenylyl cyclase and reduce steroidogenesis in mice ¹⁸; however, the precise role of PLC and its downstream signals in regulating ovarian steroidogenesis remains uncertain.

Src and Extracellular-regulated kinases (ERKs)

Activation of Gβγ by G protein-coupled receptors activates Src, which in turn triggers the Ras/Raf/MEK/MAPK/ERK pathway. For example, stimulation of LH and/or FSH receptors can enhance extranuclear receptor signaling kinase (ERK) activity ^{19, 20}, with ERK activation subsequently increasing StAR expression in some steroidogenic cell lines (e.g., Y1 adrenal cells ²¹). In contrast, Src and MAPK signaling inhibit steroid production in some cultured theca cells by decreasing activity of CYP17 and other steroidogenic enzymes ²², with some studies suggesting that lower MAPK signaling may be contributing to elevated CYP17 activity and subsequent androgen production in PCOS ²³. Given these varying observations, more experiments in different models are necessary to clearly elucidate the physiologic role of the LH-induced Src and MAPK signaling in the regulation of ovarian steroidogenesis.

Insulin-like Growth Factors (IGFs)

Several studies have implicated members of the IGF family as promoters of ovarian steroidogenesis ^{9, 24, 25}. These growth factors appear to signal through the IGF1 receptor, and, depending upon the species, have been shown to increase the levels and activities of several steroidogenic enzymes, including CYP17, CYP19 (aromatase), and CYP11A1 (side-chain cleavage enzyme). Furthermore, IGF signaling has been shown to increase LH receptor mRNA transcription, and then in turn to enhance LH-induced upregulation of StAR mRNA transcription ^{24, 26}. Details regarding intermediate IGF receptor-induced signaling pathways that regulate these downstream effects are still unclear; however, cAMP and ERK signaling may be involved ^{9, 27}.

Epidermal Growth Factor (EGF) Receptor

Another potentially important signaling molecule that is activated downstream of the LH receptor is the EGF receptor. G protein-coupled receptors (GPCRs) are known to activate EGF receptor signaling by several mechanisms ^{28, 29}. For example, stimulation of the Gα_s-coupled β2-adrenergic receptor activates membrane-bound matrix metalloproteinases (MMPs, see below) that cleave heparin-bound EGF (HB-EGF) from the cell surface to auto-activate the EGF receptor ³⁰.

Interestingly, several studies have demonstrated that EGF receptor signaling can promote steroidogenesis in gonadal cells, including Leydig cell lines and ovarian follicles ^31–33. Like LH, EGF appears to be promoting steroidogenesis primarily by activating StAR. In Leydig cells, EGF stimulates StAR production, while in ovarian follicles EGF instead appears to activate StAR post-translationally; perhaps through phosphorylation ¹³.

Since activation of the EGF and LH receptors similarly increase StAR activity and steroidogenesis, then perhaps the EGF receptor is signaling downstream of the LH receptor. In support of this hypothesis, EGF receptor activation appears necessary for gonadotropin-induced steroid production in pre-ovulatory follicles, as inhibition of the EGF receptor kinase completely abrogates LH-induced progesterone production ¹³. Intriguingly, inhibition of membrane matrix metalloproteinase (MMP)-mediated cleavage of membrane-bound EGF moieties also abrogates LH-induced steroidogenesis in ovarian follicles, suggesting the following model (Fig. 1): through yet unknown mechanisms, LH triggers activation of one or more MMPs in theca and possibly mural granulosa cells, which in turn triggers the release of membrane-bound EGF molecules, including possibly HB-EGF, amphiregulin, and epiregulin ³⁴. These soluble EGFs in turn function in both an autocrine and paracrine fashion to activate EGF receptors on theca and granulosa cells, respectively, resulting in increased StAR activity and subsequent steroid production.

Notably, as in ovarian follicles, normal LH-induced steroidogenesis in Leydig cells also depends upon activation of the EGF receptor ¹³, suggesting that EGF receptor signaling is critical for gonadotropin-mediated steroid synthesis in any gonadal tissue. In contrast, however, LH-induced steroid production in Leydig cells does not appear to require activation of MMPs. These differences may reflect the greater need for paracrine signaling in the ovary, where multiple cell types are required for steroidogenesis. Furthermore, these observations suggest that targeted inhibition of ovarian MMP activity might prove to be an effective means of specifically regulating ovarian steroid production in diseases of excess steroidogenesis, such as PCOS (see below).

Physiologic Steroid Signaling in the Ovary (Table 1)

Table 1.

Potential Local Actions of Ovarian Steroids

Steroid	Effect
Progesterone	follicle rupture, ovulation, oocyte maturation
Androgen	oocyte maturation?, FSH sensitizer, follicle growth
Estradiol	granulosa cell growth, modulation of FSH actions, oocyte maturation?, ovulation

Steroid concentrations in the ovary

Ovarian steroid levels vary depending upon the day on which they are measured, with androgen, estrogen, and progesterone levels rising during ovulation, and estrogen and progesterone levels reaching even higher levels during the luteal phase. In fact, follicular fluid at the time of ovulation has been noted to contain micromolar amounts of steroid ³⁵. What then prevents these saturating amounts of steroid from constantly signaling via their cognate receptors, which bind ligand in the nanomolar range? One possible explanation for this conundrum may be that, due to steroid metabolism in target cells, the intracellular steroid concentrations are far lower than the extracellular amounts. Alternatively, follicular fluid contains high amounts of sex hormone binding globulin, as well as other proteins that could potentially sequester steroids ^36–38; thus, while the total steroid content in follicular fluid may be high, the free, or bioavailable, steroid amounts may be relatively low.

Progesterone Actions

Direct measurement of ovarian steroid content during the LH surge in many different mammals reveals a large and rapid rise in progesterone production ^{13, 35, 39}, suggesting that progesterone may play an important role at the time of ovulation. Accordingly, progesterone signaling in the ovary seems to be essential for normal gonadotropin-induced ovulation, as injection of the PR antagonist RU486 blocks ovulation in rodents ⁴⁰, and knockout mice lacking functional PRs, fail to ovulate in response to gonadotropins ⁷. Histology in these PR null mice reveals normal development of intraovarian follicles to the tertiary follicle stage, with luteinization of granulosa cells and the production of oocytes that are competent for in vitro fertilization. In fact, the major defect appears to be the complete elimination of LH-induced follicle rupture. Additional studies have revealed the abnormal expression of ADAMTS-1 and cathepsin L, two proteases known to be involved in follicle rupture ⁴¹, indicating that trascriptional targets of the PR may be involved in the progesterone-mediated signaling cascade that triggers ovulation.

Estradiol Actions

Estradiol signals primarily via two classical receptors: ERα and ERβ. Granulosa cells mainly express ERβ, suggesting that this receptor may regulate estradiol-mediated stimulation of granulosa cell growth and modulation of FSH action. Accordingly, follicles from ERβ knockout mice have a diminished response to exogenous estradiol, with a reduced cumulus cell mass and a decreased rate of ovulation ². This defect in granulosa cell growth results in reduced female fertility, with fewer and smaller litters than wild-type mice despite normal sexual behavior.

In contrast to ERβ, ERα is expressed in every location of the hypothalamic/pituitary/gonadal axis, and ERα null mice are completely infertile. Interestingly, early ovarian development is unimpaired in the ERα knockout mice, and primordial and primary follicles are indistinguishable from those of WT females. However, concomitant with elevated serum gonadotropin levels, follicular development is markedly accelerated in the null relative to wild-type mice around day 20 of life. Follicles develop to the late antral stage but fail to progress to final maturation, often containing hemorrhagic and atretic cysts ¹. This pathology may in part be explained by a lack of negative feedback by estradiol on the pituitary, leading to excess gonadotropin production and subsequent signaling in the follicular cells.

Similar to the ER knockout mice, elimination of CYP19 aromatase expression results in excess gonadotropin secretion due to the lack of estradiol-mediated negative feedback ⁴². The hyperstimulation of ovaries by high gonadotropin levels results in the failure to ovulate. In fact, human females with CYP19 deficiency usually have ovarian failure with multicystic ovaries and mild virulization, most likely due the inability to metabolize androgens ⁴³.

Androgen Actions

Theca cells are stimulated by LH to produce androgens, which then act as paracrine regulators within the ovarian cell types. One function of androgens is to modulate FSH signaling by acting through ARs on granulosa cells to amplify cAMP levels ⁴⁴. Interestingly, AR mRNA and protein levels are appear to be down-regulated during late preovulatory development, which may serve to desensitize granulosa cells until exposure to the LH surge ⁴. The importance of the AR in ovarian function has been further examined in AR null mice, where females are subfertile with reduced litter size ^{5, 45}. Interestingly, older females lacking the AR develop premature ovarian failure ⁶, further implicating the AR as an important mediator of normal follicle development.

Nongenomic Steroid-Mediated Oocyte Maturation

One critical steroid-mediated process that is essential for normal female fertility is oocyte maturation, or re-entry into meiosis. Oocytes within the ovary are arrested in prophase I of meiosis until the gonadotropins stimulate follicular growth and development, which then triggers the resumption of meiosis up to metaphase II. Gametes are subsequently again held in meiotic arrest until fertilization, when meiosis is completed.

Oocyte maturation is best studied in Xenopus laevis, where ovarian steroids promote this process in a transcription-independent (nongenomic) fashion. In Xenopus, androgens appear to be the primary physiologic mediators of maturation in vivo, signaling via their classical ARs. In fact, blockade of androgen production significantly reduces gonadotropin induced oocyte maturation and delays ovulation ⁴⁶. Interestingly, selective androgen receptors modulators, or SARMs, have been characterized that bind to the AR and specifically regulate genomic versus nongenomic AR-mediated signaling in Xenopus laevis ⁴⁷.

The process of steroid-mediated oocyte maturation appears to be conserved in mammals, as androgens, estrogens, and progestins trigger maturation of both isolated mouse oocytes held in meiotic arrest by phosphodiesterase inhibitors ⁴⁸ and oocytes maintained in meiotic arrest by surrounding follicular cells ¹³. As in frogs, steroids appear to promote maturation via their respective classical receptors ¹³; however, since progesterone is the most abundant steroid during ovulation, it is likely the primary steroid regulating mouse oocyte maturation in vivo. Notably, in contrast to frogs, steroid-production is sufficient but not necessary for oocyte maturation in mice ^{49, 50}, suggesting that higher vertebrates have developed redundant systems to trigger this important process. One hypothesis is that steroids play an important role in selecting which oocyte will mature first in mammals, as the dominant follicles have the highest steroid concentrations ⁵¹. Importantly, regardless of their role in regulating normal oocyte maturation, excess ovarian steroid production is potentially problematic, as it could lead to unregulated oocyte maturation and subsequent infertility (see below).

Pathologic steroid signaling in the ovary – Polycystic Ovarian Syndrome

Polycystic ovarian syndrome (PCOS) is a major cause of infertility, with approximately 5–10% of all reproductive age women being affected. It is associated with high ovarian androgen production and anovulation; however, many patients with PCOS also have features of the metabolic syndrome, including obesity, hyperlipidemia, and insulin resistance ^52–54. The architecture of the ovaries in PCOS patients is abnormal, consisting of multiple irregular-sized follicles, many of which have degenerated and contain nonfunctioning oocytes. This ovarian anatomy is consistent with unregulated follicular growth resulting in the absence of dominant follicle formation that is necessary for normal ovulation.

The pathophysiology of PCOS is not well understood, and debate exists over whether insulin or androgen actions are the primary cause of the ovarian abnormalities ^54–57. In favor of insulin signaling being the primary defect is: 1) Insulin promotes follicle growth and steroidogenesis in vitro ⁵⁸. 2) Fertility in patients with PCOS and insulin resistance is often restored by insulin sensitizers that lower plasma insulin levels ^59–64.

Alternatively, other evidence points to excess androgens being responsible for the ovarian pathology in PCOS patients. First, polycystic ovaries and infertility are found in women with other forms of androgen excess that often are not accompanied by insulin resistance, including exogenous androgen use (e.g., female transgender patients and women on anabolic steroids ^{65, 66}) or excess adrenal androgen production (e.g., congenital adrenal hyperplasia ⁶⁷). Second, androgen receptor (AR) antagonists such as flutamide restore ovulation in certain subsets of patients ^68–71. Finally, physiologic levels of androgens and androgen receptors are important for normal follicle development ^{6, 72}, and therefore might enhance growth when in excess.

The following model can been used to reconcile some of these complex issues regarding PCOS and insulin resistance: normal insulin levels may stimulate baseline ovarian androgen production, which in turn participates in regulating follicle development. In contrast, excess insulin may promote increased ovarian androgen production ⁵⁸, resulting in unregulated follicle growth with an absence of dominant follicle formation and subsequent ovulation.

If elevated androgen signaling in the ovary contributes to the ovarian pathology in PCOS, then AR antagonists should be helpful in treating this disease. As mentioned, both flutamide and spironolactone have indeed been used successfully to improve fertility in PCOS ^69–71; however, the efficacy of these drugs is limited due to their low potency. As an alternate approach, rather than use weak AR antagonists to block androgen effects, targeted reduction of ovarian androgen production might attenuate excess androgen signaling and restore fertility in women with PCOS. One possible means of decreasing ovarian androgen production might be to block MMP or EGF receptor activity, both of which are important for ovarian steroidogenesis. Alternatively, specific inhibition of CYP17 lyase might reduce androgen production without affecting adrenal cortisol synthesis in the adrenal gland ⁷³. Further translational studies linking laboratory research to the infertility clinic will allow these types of studies to be performed, and may lead to new and exciting advances in the field of female reproductive biology.