The Role of Growth Factors in Ovarian Function and Development

Ovarian folliculogenesis is a dynamic process marked by exponential expansion and differentiation of the granulosa cells, maturation of the oocyte, and neovascularization (Figure 1 to follow).

Although the central roles of gonadotropins and of gonadal steroids in this explosive agenda are well accepted, the variable fate of follicles within the same ovary suggests the existence of additional intraovarian modulatory systems.¹ Stated differently, it is presumed that gonadotropin action is 'fine tuned' in situ, thereby accounting for observed differences in the rate and extent of development of ovarian follicles. Alterations in gonadotropin secretion cannot adequately explain the initiation and arrest of meiosis within the oocyte, the acquisition of follicular dominance, or the failure of follicular development, which leads to atresia. It is likely that the earlier stages of follicular growth, generally considered to be gonadotropin independent, may be controlled by intraovarian signaling.

The concept of local gonadal regulators originated when the embryology of the ovary was the subject of intense scrutiny. Gonadal differentiation was proposed by Witschi to result from the interaction of two morphogenic substances called cortexone and medullarine, the first of which was thought of as the stimulator of ovarian development and the latter as the promoter of testicular growth.² Although multiple other contributors must undoubtedly be acknowledged, the notion of intraovarian regulators was promoted with special vigor by the late Cornelia Post Channing, whose pioneering experiments ushered in contemporary molecular endocrinology as it applies to ovarian physiology.³

Among potential novel intraovarian regulators, growth factors, cytokines, and neuropeptides have been the subject of increasingly intense investigation. Most of these agents are not expected to act in the traditional endocrine fashion because of their local intraovarian generation (as opposed to circulatory-derived influences emanating from distant endocrine glands). Speculation favors the notion that a host of putative intraovarian regulators may engage in subtle in situ modulation and coordination of growth and function of the varied follicular cell types: oocytes, granulosa, theca, and vascular epithelium (Figure 2 to follow).

In this capacity, a given putative intraovarian regulator may modulate the replication or cytodifferentiation of a developing ovarian cell, acting in its own right or as an amplifier-attenuator of gonadotropin action. Such putative intraovarian regulators may also be concerned with intercompartmental communication, allowing tighter linking of different cellular populations. For example, a growing body of evidence suggests that granulosa cell-derived modulators may regulate the adjacent theca-interstitial cell compartment in the interest of coordinated follicular development. In doing so, the granulosa cell may exert some control over its own destiny, in that it may regulate the very inflow of androgenic substrate from the neighboring theca. Together, gonadotropins, steroids, and locally derived peptidergic principles form a triad, which modulates the growth and differentiation of ovarian follicles (Fig. 3). According to contemporary views, potential intraovarian communication is mostly paracrine or autocrine in nature. Paracrine communication involves local diffusion of regulators from producer cells to distinct target cells within the same organ. This is a heteroregulatory phenomenon that could allow for intercompartmental communication, providing a tighter linkage of different cellular populations. In the ovary, the ability of increasing numbers of granulosa cells to produce estrogen depends on the concomitant ability of the thecal layer to provide the proper amounts of androgenic substrate. The granulosa cell, in the interest of efficient coupling, may elaborate substances (e.g. insulin-like growth factor-I [IGF-I] inhibin, activin) that could alter the function of the neighboring theca.

Fig. 3 Modulators of ovarian follicular growth and development: the regulatory triad

The other type of cellular communication, autocrine regulation, involves the action of a regulator on surface receptors at its cell of origin. This is a self-regulatory phenomenon wherein a single cell type modulates its own activity. In the ovary, granulosa cells elaborate substances such as IGF-I and activin that can alter granulosa cell function. Whereas steroids may be exerting intracrine (i.e. regulation within the cell of origin) effects, there is no evidence for juxtacrine (i.e. contact-dependent regulation between immediately adjacent cells) effects in the ovary.

To qualify as a bona fide intraovarian regulator, the putative agent needs to meet the minimal criteria of local production, local reception, and local action. Some evidence of indispensability to in vivo ovarian function needs to be provided. For the most part, few of the putative intraovarian regulators under study (Table 1) have satisfactorily met all of the previously described criteria (i.e. IGF-I, activin). Accordingly, the information provided later can be viewed as a prelude to what the future holds. Undoubtedly, additional information will become available with respect to the putative intraovarian regulators under consideration. It is equally certain that novel candidates will be added to this preliminary list, requiring modification of current views.

Table 1. Established and putative intraovarian regulators

  Insulin-Like Growth Factor System
  IGF-I
  IGF-II
  IGF binding proteins

  Inhibin/Activin Systems
  Inhibin
  Activin
  Follistatin

  Interleukin-1 System
  Interleukin-1
  Interleukin-1 receptor antagonist
  IL-1 binding protein (IL-1 receptor type II)

  Other Growth Factors
  EGF/TGFα
  TGFβ1, TGFβ2
  NGF
  aFGF, bFGF
  VEGF
  TNFα

  Other Peptidergic Factors
  Ovarian renin angiotensin system
  VIP
  Oxytocin
  Endothelin

The following sections describe a select group of putative intraovarian regulators reflecting different modes of action. The principle action of each regulator is briefly listed in Table 2.

Table 2. Principal actions of intraovarian regulators

  Insulin-like growth factor-I
  Follicle-stimulating hormone (FSH) amplification
  Follicular growth
  Follicular selection

  Transforming growth factor-α
  Follicular maturation
  Oocyte maturation
  Cellular differentiation
  Potentiation of gonadotropin action
  Regulation of apoptosis

  Transforming growth factor-β1
  Follicular rupture inhibition
  Follicular differentiation

  Basic fibroblast growth factor
  Apoptosis inhibition
  Regulation of folliculogenesis

  Activin
  Oocyte maturation
  Follicular differentiation
  Early embryogenesis
  Regulation of steroidogenesis

  Interleukin-1 (see also Fig. 4)
  Ovulation induction
  Glycolysis
  Glucose transport

  Tumor necrosis factor-α
  Inhibits steroidogenesis
  FSH antagonist
  Induces apoptosis/luteolysis
  Ovulation inhibition

Fig. 4 Intraovarian interleukin-1 as a mediator of gonadotropin action

A 70-amino acid polypeptide, IGF-I plays a variety of metabolic and endocrine roles, not the least of which is the promotion of linear skeletal growth. In keeping with its ubiquitous distribution, IGF-I is also known to serve a variety of autocrine or paracrine tissue-specific functions to suit the needs dictated by the tissues in question. In this respect, the ovary is but one example of many exemplifying the general concept of intraorgan regulation.⁴

A large body of information now strongly supports the view that the ovary is a site of IGF-I production, reception, and action (Fig. 5). Whereas the rat granulosa cell appears to be the only cellular site of IGF-I gene expression,^{5, 6} the granulosa^{7, 8} and the theca-interstitial cells^{9, 10} possess specific receptors for this peptidergic ligand. The mouse intraovarian IGF-I system is generally comparable to that of the rat, although they differ in several aspects.^{11, 12} These observations suggest that IGF-I may engage in intercompartmental communication in the interest of coordinated follicular development. IGF-I hormonal action appears subject to further modulation through the local elaboration of low-molecular-weight binding proteins (IGFBPs), the role and regulation of which are receiving increasing attention. Whereas the main IGFBP is IGFBP-3, being up-regulated by GH, other IGFBPs bear important roles in reproductive endocrinology. For instance, IGFBP-1 being down-regulated by insulin, has an important role in the pathophysiology of PCOS, by increasing the free, biological active IGF-I, augmenting the androgen generation in the theca layer.¹³ The discovery of IGFBP-4 mRNA in early-stage atretic follicles raises the intriguing possibility that depletion of IGF action may be necessary for the onset of the atretic process.¹⁴ Although multiple ovarian actions have been ascribed to IGF-I, its main role appears to be the amplification of gonadotropin action in theca-interstitial and granulosa cells. All markers of follicle-stimulating hormone (FSH) induction (e.g. production of progesterone inhibin, luteinizing hormone binding) are enhanced by IGF-I. Optimal gonadotropin hormonal action is contingent on the prior availability of granulosa cell-derived IGF-I and the consequent amplification of the gonadotropic signal. Given a hypothetical IGF vacuum created by excess exogenous IGF binding proteins, intrinsic FSH hormonal action proves to be relatively modest (Fig. 4). In contrast, given IGF-replete circumstances, FSH hormonal action in toto may be composed of a modest intrinsic component complemented by a substantial synergistic component.¹⁵ IGF-I has an obligatory role in granulosa cell replication in all species tested.¹⁶ Further consideration must be given to the possibility that there are two distinct types of granulosa cell related to their proximity to the oocyte.¹⁶ In addition, there appears to be an association between increased bioavailability of IGF-I in follicular fluid and selection of the dominant follicle.^{17, 18}

Fig. 5 The intraovarian insulin-like growth factor-I system

Fig. 6 Enhancing effect of insulin-like growth factor-I on follicle-stimulating hormone-stimulated progesterone accumulation

At the clinical level, ovarian IGF-I may have a bearing on the puberty-promoting effect of growth hormone. An association appears to exist between isolated growth hormone deficiency and delayed puberty in rodents and human subjects, a process reversed by systemic hormone replacement therapy. Given that ovarian IGF-I and its receptor may be growth hormone dependent, it is tempting to speculate that the ability of growth hormone to accelerate pubertal maturation in part may be caused by the promotion of ovarian IGF-I production and reception with the consequent local potentiation of gonadotropin action.

Clear evidence for the central role of IGF-I in reproductive physiology has been gained from gene knockout technology. In the mouse, targeted null mutation of the Igf1 gene, encoding IGF-I, results in infertility secondary to failure to ovulate even after administration of gonadotropins.¹⁹ Given the IGF-I primary action of FSH amplification, further efforts have been made to elucidate its mechanism of action in that regard.^{20, 21, 22} IGF-I leads to increased estradiol synthesis and proliferation of granulosa cells, enhanced LH-induced androgen-synthesis in theca cells and increased inhibin-, activin- and follistatin secretion from granulosa cells.²³

In human dominant follicles IGF-II receptor is expressed in both granulosa cells and theca cells, while IGF-II is expressed extensively in granulosa cells.²⁴ In addition, IGF-II is expressed in the corpus luteum, where it coordinates vessel maintenance and angiogenesis.²⁵ IGF-II increases in the presence of FSH, LH receptor synthesis in theca-interstitial cells and granulosa cells from antral follicles,^{26, 27} as well as proliferation and steroidogenesis in granulosa cells.²⁶

Growth hormone receptor (GH-R) is expressed in the rat ovary together with GH binding protein (GHBP) that regulates the bio-availability of GH.^{28, 29} GH and GH-R are detected in oocytes and granulosa cells in early developing human follicles.³⁰ However, it is still not clear whether systemic GH binds to GH-R in the ovary, or the ovary itself produces GH.³¹ It is suggested, that in the ovary GH acts directly through ovarian GH-R, or indirectly via IGF-I and IGF-II.³¹ GH has a stimulatory effect on the formation of secondary follicles.³² GH-R and GHBP deficient mice have strongly reduced numbers of primary, secondary and antral follicles, but an elevated number of primordial follicles and a significant increased number of atretic follicles.³³ In addition, GH is responsible for the development and maintenance of sensitivity to gonadotropin.^{34, 35} GH increases IGF-I secretion from theca cells and IGF-I stimulates proliferation of granulosa cells and steroidogenesis.³⁶ On the other hand, it has been shown that GH also directly enhances estradiol production in cultured granulosa cells.³⁷ GH promotes nuclear and cytoplasmic maturation and improves the developmental capacity of oocytes.^{38, 39}

Purified on the basis of its ability to stimulate precocious eyelid opening and tooth eruption in newborn mice, epidermal growth factor (EGF) was initially found in male mouse submaxillary glands and later in human urine as urogastrone. Mature EGF comprises a single polypeptide chain of 53 amino acids displaying three internal disulfide bonds. Originally thought to have a limited range of tissue expression, in situ hybridization analysis of sections of whole newborn mice indicate that RNA complementary to cloned EGF probes may be present in a large variety of tissues.

Transforming growth factor-α (TGF-α), a structural analog of EGF, is a single-chain, 50-amino acid polypeptide capable of binding to an apparently common EGF/TGF receptor. EGF and TGF recognize the same cellular receptor, and they are apparently equipotent in most systems studied. EGF may be the adult form of the embryonic growth factor TGF. TGF is a member of a family of polypeptides best known for their ability to produce an acute, albeit reversible, phenotypic transformation of normal mammalian cells. TGF can be defined operationally by its ability to stimulate anchorage-independent growth in soft agar of cells, which are otherwise anchorage dependent.

At the level of the ovary, EGF exerts potent regulatory effects on granulosa cell proliferation and differentiation.^{40, 41, 42} These effects of EGF presumably are mediated by specific cell membrane receptors, the existence of which has been demonstrated on bovine, ovine, and murine granulosa cells.⁴³ However, the identity of the endogenous ligand occupying the receptor in question under in vivo conditions remains uncertain. Norris et al., 2010⁴⁴ demonstrated that EGF receptor kinase contributes to LH-induced meiotic resumption of oocytes through the closure of gap junctions and through a decrease in follicle cGMP.

EGF-like growth factors

It has been shown, that the binding of LH to G-a-coupled LH receptors on the outer theca cells and mural granulosa cells leads to matrix metaloproteinase (MMP)-mediated release of EGF-like growth factors from the cell surface.⁴⁵ Subsequently, the released EGF-like growth factors bind to EGF receptors located on cumulus cells. Such EGF-like growth factors are amphiregulin, betacellulin and epiregulin.⁴⁶ It has been suggested, that amphiregulin is important for oocyte maturation and cumulus cell expansion.⁴⁷

Androgens and progestins promote oocyte maturation through steroid receptors.⁴⁸ LH-induced, MMP-mediated,  release of EGF-like growth factors leads to activation of EGF receptors on cumulus granulosa cells and hence to phosphorylation of steroidogenic acute regulatory protein (StAR), resulting in subsequent up-regulation of steroidogenesis.⁴⁹

TGF, like EGF, proved to be a potent inhibitor of gonadotropin-supported granulosa cell differentiation. TGF has been localized to the theca-interstitial cell compartment,⁵⁰ thereby raising the possibility that theca-interstitial cell-derived TGF may exert paracrine effects at the level of the adjacent granulosa cell. Theca-interstitial cell-derived TGF may also engage in autocrine effects.⁵¹ It is tempting to speculate that TGF of theca-interstitial cell origin may orchestrate follicular activities at the granulosa and theca-interstitial cell level (Fig. 7). However, because TGF has also been shown to suppress gonadotropin-supported theca-interstitial cell differentiation,⁵¹ the possibility of an autocrine mode of action cannot be excluded. Further evidence supports the possibility that TGF may also be expressed by other compartments of the ovary (e.g. granulosa cells, oocytes). In humans, the expression pattern may be age and cycle dependent⁵² and may have a role in ovarian embryogenesis.⁵³

Fig. 7. The intraovarian epidermal growth factor/transforming growth factor-α system. EGF, epidermal growth factor; TGF, transforming growth factor; FSH, follicle stimulating hormone.

TGF is also involved in the process of follicular apoptosis, which is central in maintaining a balance between cell proliferation and demise. Treatment of cultured granulosa⁵⁴ or theca-interstitial cells⁵⁵ with TGF inhibits the spontaneous onset of apoptotic DNA cleavage.

Transforming growth factor-β (TGF-β) superfamily is a group of about 35 proteins involved in pre- and postnatal physiological processes.⁵⁶  Members of this superfamily are expressed by oocytes and ovarian somatic cells in key developmental stages.^{57, 58, 59, 60, 61} Throughout life ovarian follicles leave the resting pool to join the growing pool. The precise mechanism of follicular recruitment is not fully elucidated, however, members of the TGF-β family are involved in the process. Bone morphogenetic proteins (BMP) 7 and 4 promote primordial to primary follicle transition.⁶² Similarly, anti-Mullerian hormone (AMH), another member of the TGF-β superfamily, is involved in initiation of primordial follicle growth.⁶³ Progression of primary follicles to early antral stage is enhanced by growth and differentiation factor-9 (GDF-9) and BMP-15 of oocyte origin, activins of granulosa origin, and BMP-4 and BMP-7 of thecal origin.⁶¹ Antral follicle growth and follicle selection mechanism involves the inhibin-activin system.⁶⁴ Activin, TGF-β, and several BMPs exert paracrine actions on theca cells to attenuate LH-dependent androgen production in small to medium size antral follicles.⁶⁵ Dominant follicle selection is influenced by changes in intrafollicular activins, GDF-9, AMH, and several BMPs. Activin plays a positive role in oocyte maturation, while inhibin upregulates LH-induced androgen secretion to sustain estradiol biosynthesis during the pre-ovulatory phase.

In human ovaries GDF-9 promotes the growth of early pre-antral follicles.⁶⁶ GDF-9 inhibits FSH-induced steroidogenesis in pre-ovulatory follicles and promotes progesterone production in cumulus cells through enhancing the expression of an intrinsic prostaglandin-E2/EP2 receptor signaling pathway.⁶⁷ Furthermore, GDF-9 promotes follicular survival during the transition to the antral stage by suppressing granulosa cell apoptosis and follicular atresia.⁶⁸ During the latter stages of folliculogenesis cumulus cells require GDF-9 in order to support glycolysis and sterol biosynthesis prior to the LH-surge.⁶⁹ It has been shown that GDF-9 protects granulosa cells from apoptosis through the activation of the PI3K/Akt pathway in pre-antral follicles ⁶⁸. In addition, GDF-9 regulates granulosa cell mitosis through both Smad-dependent and independent pathways.⁷⁰ Mutations in human GDF-9 contribute to ovarian insufficiency in women through defective GDF-9 production and/or activation.⁷¹ It has been shown that women with mutations leading to the activation of latent human GDF-9 reach menopause before the age of 35 years. These women have elevated FSH and LH levels and atrophic ovaries devoid of follicles. It is suggested that mutations in the latency-associated GDF-9 prodomain may contribute to premature ovarian failure through an increase in the number of growing follicles leading to a premature depletion of the ovarian reserve.⁷¹

Anti-Mullerian hormone

AMH is a member of the TGF-β superfamily.⁷² AMH is solely expressed in granulosa cells of small growing follicles. After the initiation of primordial follicle growth, AMH is expressed in granulosa cells and remains expressed until the small antral stage. Then, FSH leads to a decrease in AMH expression and is almost absent in granulosa cells during the FSH-dependent follicular growth.⁷³ In contrast to this, AMH expression sustains in cumulus cells of pre-ovulatory follicles.⁷⁴ AMH receptor II shows a similar expression pattern as AMH itself⁷⁵ and is also expressed in theca cells.⁷⁶

During early folliculogenesis AMH inhibits primordial to primary follicle transition and progression to the antral stage.⁷⁷

In human granulosa cells AMH inhibits the effects of FSH.⁷⁸ It has been shown, that AMH inhibits FSH-dependent aromatase synthesis and estradiol production in human granulosa cells.⁷⁹ AMH probably keeps the follicle relatively insensitive to FSH until the follicle reaches a certain grade of maturity.⁸⁰ In accordance to this, it has been shown that in AMH-knockout mice, more small antral follicles and also the larger, normally not FSH-responsive pre-antral follicles were recruited in the absence of AMH.⁸¹ It can be concluded that with decreasing AMH expression the FSH-sensitivity of the follicle increases and therefore the follicle can be recruited to enter the pool of follicles that may become dominant (Figure 8 to follow).

Basic fibroblast growth factor (bFGF), a 146-amino acid polypeptide, is a mitogen for a wide variety of mesoderm-derived and neuroectoderm-derived cells. Its complete isolation and characterization has been accomplished from various organs; an amino terminally truncated form lacking the first 15 residues was identified in the ovarian corpus luteum.⁶⁵ Although the physiologic relevance of bFGF to ovarian function remains under investigation,⁸² several lines of evidence suggest that bFGF may play a central role in supporting the growth and development of the granulosa-luteal cell. Basic FGF constitutes the main mitogenic factor isolated from crude extract and has previously been shown to stimulate the replicative lifespan of cultured granulosa cells of bovine, porcine, rabbit, guinea pig, and human origin.^{83, 84, 85} Because ovarian bFGF expression was not considered to be of granulosa cell origin,⁸⁶ whereas FSH induces functional receptors for bFGF in the granulosa cells,⁸⁷ it is tempting to speculate that locally produced bFGF⁸⁸ may play autocrine or paracrine regulatory roles at or adjacent to its sites of synthesis. In so doing, it may participate in the differentiation and replication of the developing granulosa cell.⁸⁹

Basic FGF is involved in early development of the human reproductive tract⁹⁰ and partakes in suppression spontaneous onset of apoptosis.⁹¹ The latter may be associated with the ability of progesterone to maintain granulosa cell viability.⁹² In addition to granulosa cells, bFGF also inhibits apoptosis of the ovarian surface epithelial cells, acting in both sites on its own receptor.⁹³ Basic FGF can be identified in a host of ovarian components, including granulosa cells, oocytes, follicular basement membrane, and surface epithelial cells.^{94, 95, 96}

Activin is a 24-kD protein with structural homology to TGF-β1. It was discovered during the purification of inhibin and found to be a dimer of the β subunits of the heterodimeric inhibin molecule.⁹⁷ Activin was concurrently discovered as capable of differentiating erythroleukemia cells⁹⁸ and inducing mesoderm formation.⁹⁹ Its presence in a variety of cell types suggests that it may regulate growth and differentiation in other tissues as well.⁹⁷

Activins play a role in the local regulation of ovarian function. Acting through a set of receptors, postulated to be membrane-bound serine/threonine kinases,^{100, 101} activin alters the function of granulosa and theca-interstitial cells. For instance, activin treatment of cultured granulosa cells from immature follicles increases FSH-supported estradiol production, inhibin production, and FSH and luteinizing hormone binding. Activin may maintain the immature follicle during the period of declining FSH levels, which is induced by its partner inhibin. In contrast to its action on immature granulosa cells, activin decreases progesterone production by mature granulosa cells from preovulatory follicles.¹⁰² Based on these observations, Findlay and coworkers¹⁰³ proposed an autocrine role for activin as a suppressor of spontaneous luteinization. Paracrine actions of activin are also a possibility because activin reduces luteinizing hormone-induced androstenedione production by cultured theca-interstitial cells.^{104, 105}

Throughout follicle development the balance between activin and inhibin expression shifts. Activins are especially expressed in primary and antral follicles, while inhibins are expressed mainly in larger more developed follicles.¹⁰⁶ Activin plays an essential role in primordial follicle assembly and in the establishment of the size of the primordial follicle pool. It has been suggested that during the time of follicle assembly the local activin concentrations determine the size of the ovarian follicle pool.¹⁰⁷ As such it has been demonstrated that the administration of activin to neonatal mice at the time of germline cyst-breakdown led to an increase in the number of germ cells and pre-granulosa cells and also to an increase in the size of the primordial follicle pool.¹⁰⁸ In accordance to this observation, Lei et al., 2010¹⁰⁹ found that FSH treatment to a neonatal mouse ovarian culture model increased proliferation of pre-granulosa cells, expression of activin b and bB subunits and oocyte survival, and thus promoted overall primordial follicle formation. In addition, it is assumed that activin has the capacity to suppress early follicle growth.¹⁰⁷ As such it was demonstrated that the presence of activin secreting secondary follicles inhibited the growth of pre-antral follicles in an in vitro co-culture model.¹¹⁰ Furthermore, activin increases LH induced androgen synthesis in small and medium sized antral follicles¹¹¹ and promotes oocyte maturation and growth of pre-antral follicles.¹¹² In this regard it has been shown that activin A alone, and in combination with FSH, promotes granulosa cell proliferation in granulosa cells of pre-antral follicles and oocyte growth.¹¹³

Activin stimulates the ERa promoter and estrogen receptor activity. On the other hand, estrogen suppresses activin b subunit and activin bB subunit. With the progression of folliculogenesis, activins may become inhibitory to steroidogenesis.¹¹⁴ In accordance to this, it has been demonstrated that in late tertiary follicles activin inhibits FSH-mediated aromatase activity.¹¹⁵ Activin has an anti-luteinization effect on granulosa cells. In this context, activin upregulates FSH-R and HSD11b2, downregulates LH-R, blocks the hCG-induced upregulation of StAR and downregulates ERa expression.¹¹⁶ In luteinized human granulosa cells activin suppresses progesterone production and the expression of CYP11a1, HSD3 b and CYP19.¹¹⁷ It is assumed that a reduction in activin signaling is essential for luteinization and the formation and maintenance of the corpus luteum. During luteolysis activin enhances MMP-2 expression, which can be reversed by follistatin.¹¹⁸ Activin receptors are expressed in the corpus luteum,¹¹⁹ while the level of follistatin decreases in regressing corpus luteum, thus resulting in increased bioavailability of activin.¹²⁰

Follistatin, a glycoprotein with isoforms of 35–40 kD, was also discovered during the purification of inhibin.⁹⁷ Its ability to bind activin¹²¹ provides a possible explanation for the observation that follistatin antagonizes the in vitro actions of activin. The presence of follistatin primarily in preovulatory follicles¹²² supports the idea that blocking activin is necessary for maturation and luteinization. In the developing follicles, FSH-induced granulosa cell proliferation and mitogenesis is facilitated by activin.¹²³

The study of activin action is further complicated by the ability of its component subunits to combine with the α subunit of inhibin to form a molecule whose action in many experimental assays is diametrically opposed to that of activin. Although autocrine actions of inhibin have not been convincingly demonstrated, granulosa cell-derived inhibin can oppose the activin blockade of thecal androgen production.¹⁰⁴ Activin, follistatin, and inhibin form a complex mix of intraovarian regulators (Fig. 9).

Fig. 9 The intraovarian activin/inhibin system. LH,luteinizing hormone; hCG, human chorionic gonadotropin

Inhibin exists in two isoforms, inhibin A and inhibin B. Inhibin A consists of one a- and one bA subunit, while inhibin B comprises one a- and one bB subunit.¹⁰⁷ It has been shown that both inhibin isoforms are expressed in granulosa cells of growing follicles.¹²⁴ Inhibin A is especially produced by the dominant follicle after the preovulatory gonadotropin surge, and by the corpus luteum, whereas inhibin B is mainly secreted from the small antral follicles.^{125, 126} In the ovary inhibin antagonizes the action of activin through its competitive binding to activin receptors.¹²⁷ Inhibin enhances in theca cells the LH-induced androgen synthesis.¹²⁸ In neonatal mice the level of inhibin B drops drastically during the interval of follicle formation. This in turn lowers the inhibin to activin ratio and hence leads to sustained activin signaling in the ovary during the period of follicle formation.¹⁰⁷ Vitale et al., 2002¹²⁹ suggest that inhibin may play an important role in follicular selection, as inhibin secreted by the dominant follicle leads to apoptosis in subordinate follicles.

The corpus luteum is a major source of inhibin A during the human menstrual cycle.¹³⁰ It has been demonstrated that antibodies to the a subunit of inhibin led to a decrease in hCG-induced progesterone secretion by luteal cells (Figure 10 to follow).¹³¹

Interleukin-1 (IL-1), a polypeptide cytokine previously referred to as lymphocyte-activating factor, is predominantly produced and secreted by activated macrophages. It possesses a wide range of biologic functions and plays a role as an immune mediator.¹³² At the level of the ovary, IL-1 suppresses the functional and morphologic luteinization of cultured murine and porcine granulosa cells.^{133, 134} Exerted at physiologic concentrations (10-⁹ M), IL-1 action could not be attributed to altered cell viability. Rather, the antigonadotropic activity of IL-1 appeared to involve sites of action proximal and distal to cAMP generation. Subsequent work by Kasson and Gorospe shed additional light on the ovarian relevance of interleukins.¹³⁵ IL-1α and IL-1β augmented the FSH-stimulated accumulation of 20-dihydroprogesterone. In all cases, less IL-1β than IL-1α was required to produce a comparable effect. Other studies in the rat ovary indicate that the rat ovarian theca-interstitial cell is a site of IL-1β gene expression, the preovulatory acquisition of which is gonadotropin dependent.¹³⁶ However, immediately after follicle rupture, granulosa cells stain positive for IL-1β in immunohistochemical studies in the mouse ovary.¹³⁷ The possibility of a shift in IL-1β origin, receptor, and action to the granulosa cell compartment just before ovulation cannot be excluded.

Although the relevance of IL-1 to ovarian physiology remains a matter of study, it is tempting to speculate that IL-1 could be involved in mediation of gonadotropin action and in the luteinization process (Fig. 4). Such speculation appears particularly intriguing in light of the apparent progesterone dependence of IL-1 gene expression.¹³⁸ In contrast, higher concentrations of progesterone significantly inhibit IL-1 activity.¹³⁹ Although much remains to be learned on the intraovarian cellular origin of IL-1, resident interstitial ovarian macrophages could be sites of hormonally regulated IL-1 gene expression given the reported gonadotropin dependence of their testicular counterparts.¹⁴⁰

Significant amounts of IL-1-like activity have been detected in follicular fluid.¹⁴¹ The ovarian reception of IL-1 involves the type I IL-1 receptor, whose transcripts have been identified in cultured human granulosa and theca cells.¹⁴²  Interleukin and its receptors are maximally expressed in granulosa cells and theca cells in preovulatory follicles after the gonadotropin action.¹⁴³ It was shown that IL-1 signaling occurs exclusively through the type I receptor,¹⁴⁴ whereas the type II receptor inhibits IL-1 activity by acting as a 'decoy' target for IL-1.¹⁴⁵ A growing body of evidence supports the role of IL-1 as an intermediary in the ovulatory process (Fig. 4). IL-1 is a potent stimulator of the ovarian phospholipase A₂ system,^{146, 147, 148} and of prostaglandin endoperoxide synthase-1 and -2,¹⁴⁹ both in the interest of upregulating prostaglandin biosynthesis. IL-1 is also involved in ovarian carbohydrate economy.^{150, 151,} Ovarian IL-1 signaling occurs through the type I receptor, the expression of which is stimulated with ovulation.¹⁵²In vivo models, using perfused ovaries or direct intrabursal injection, also support IL-1's central role in the process of ovulation.^{153, 154, 155}

IL-1 can induce NO production in the ovary.¹⁵⁶ In humans IL-1β can increase NO production by follicular cells after a 24 hour incubation period.¹⁵⁷ Increasing NO production by IL-1β can inhibit apoptosis in rat ovarian follicles.¹⁵⁸ In addition, it has been shown that IL-1beta inhibits estradiol production indirect via the stimulation of NO production.¹⁵⁹ Furthermore, IL-1 inhibits gonadotropin stimulated secretion by granulosa cells. This may be due to a decrease in CYP19 activity.¹⁶⁰ Altogether, it is suggested that IL-1 is involved in the intra-ovarian regulation of steroid biosynthesis.¹⁶¹

Granulosa cells synthesize IL-6¹⁶² and the IL-6 receptor is expressed in granulosa cells as well.¹⁶³ It has been shown that IL-6 reduces the expression of CYP19¹⁶³ as well as LH-R expression in granulosa cells.¹⁶⁴ It is suggested that IL-6 mediates some of its effects in the ovary through the activation of the ERK1/2, JAK/STAT and p38MAPK pathways.¹⁶⁵ As such IL-6 is a potent regulator of cumulus cell function and oocyte-cumulus cell expansion. In the corpus luteum, IL-6 inhibits hCG induced progesterone secretion from luteal cells and hence IL-6 is suggested to act as an autocrine and paracrine regulator in the corpus luteum at the time of luteolysis.¹⁶⁶

TNFα, a 157-amino acid polypeptide, was originally named for its oncolytic activity as displayed in the serum of bacillus Calmette-Guérin-immunized, endotoxin-challenged mice.^{167, 168} TNF proved capable of inducing tumor necrosis in vivo and of exerting non-species-specific cytolytic or cytostatic effects on a broad range of transformed cell lines in vitro. Although TNF was initially thought to be tumor selective, it has become clear that certain nontumor cells possess TNF receptors and that TNF may be a regulatory monokine with pleiotropic noncytotoxic activities in addition to its antitumor properties. TNF engages in the differentiation of a variety of cell types.

At the level of the ovary, TNF was found capable of attenuating the differentiation of cultured granulosa cells from immature rats.¹⁶⁹ In other studies, TNF was found to effect complex dose-dependent alterations in the elaboration of progesterone and androstenedione, but not estrogen, by explanted preovulatory follicles of murine origin. Although the ovary contains TNF mRNA,¹⁷⁰ its in vivo origins must be determined. In principle, two general possibilities are worthy of consideration. TNF may be locally derived from (activated) resident ovarian macrophages,¹⁷¹ as shown for regressing (but not young) corpora lutea. Although basal TNF activity was undetected in corpora lutea of pregnancy and pseudopregnancy, TNF activity was markedly stimulated in the presence of lipopolysaccharide.¹⁷² However, the detection of TNF activity in some luteal tissue on day 5, and the scarcity of macrophages at this stage raise the possibility that cells other than macrophages may also produce TNF in the corpus luteum. TNF may be of granulosa cell origin, as suggested by immunohistochemical studies wherein antral or atretic granulosa cells have been implicated as a possible site of TNF gene expression. Given such strong association between TNF elaboration and follicular and luteal decline, it is tempting to speculate that TNF may play a role in the still enigmatic processes of atresia or luteolysis. In this capacity, TNF of intraovarian origin may exert its effects at or adjacent to its site of synthesis, interacting with specific granulosa-luteal cell surface receptors to modulate gonadotropin hormonal action. TNF-induced luteolysis¹⁷³ by apoptosis has been well documented.¹⁷⁴ TNF induced apoptosis depends on the ceramide signaling pathway as its second messenger.^{175, 176}

It has been shown that null mutations in TNF type I receptor (TNF-RI) impaired ovarian cycling in aged females and increased pre-pubertal ovarian responsiveness to gonadotropins, while null mutations in TNF-RII did not cause any effect regarding fertility in mice.¹⁴³

TNFα is an inductor of apoptosis in granulosa cells and also of follicular atresia.¹⁷⁷ TNFα mediates apoptosis through its receptors and the downstream mediators TRAIL, TRADD and TRAF2.

It has been shown that TNF stimulates progesterone synthesis in differentiated ovaries, while in undifferentiated ovarian cells TNF inhibits steroidogenesis.¹⁷⁸

When pregnancy is established, the corpus luteum must produce progesterone to maintain the pregnancy. While TNFα and its receptors have been documented in the gravid uterus, placenta and embryo¹⁷⁹ high affinity binding sites for TNFα were demonstrated in the corpus luteum.¹⁸⁰ It is possible that locally produced TNFα plays an important role as an autocrine and/or paracrine mediator in the corpus luteum during pregnancy. Luteal TNFα may contribute to maintaining the pregnancy by stimulating the production of PGF2α and PGE2 by the corpus luteum of pregnancy,¹⁸¹ indirectly resulting in an increase in progesterone output.

Undoubtedly, future studies of the regulation of the TNF receptor and the elucidation of the in vivo source of its ligand will shed new light on the relevance of this system to the process of follicular development or demise.

Ovarian follicles produce colont stimulating factor (CSF).¹⁸² It is suggested that CSF-1 acts locally in the ovary on gonadotropin receptors.¹⁴³ As such it has been demonstrated that mutations in CSF-1 lead to dysfunctional LH secretion, impaired ovarian luteinization, decreased ovulation and reduced occurrence of antral follicles.¹⁴³

Other growth and peptidergic factors have potential physiologic relevance to folliculogenesis (Table 1). The ovary contains a complete renin-angiotensin system that may be involved with vascularization and with modulation of steroidogenesis.¹⁸³ Vasoactive intestinal peptide is also produced locally in the ovary, and it can enhance estrogen production by granulosa cells of prepubertal rats.¹⁸⁴

Nerve growth factor is another peptidergic factor whose mRNA has been detected in the ovary,¹⁸⁵ but its modulatory role in the ovary is unknown. Similarly, endothelin, a potent vasoconstrictor, influences ovarian progesterone production.¹⁸⁶

Relaxin

Relaxins are peptide hormones and bind to G-protein coupled receptors.¹⁸⁷ Relaxin is a major product of the corpus luteum during pregnancy¹⁸⁸ and is mainly produced by luteal granulosa cells.¹⁸⁹ Relaxin is also expressed in theca interna cells of antral follicles before the LH-surge.¹⁹⁰

In the ovary, relaxin binds to the G-protein coupled receptor, RXFP1,¹⁸⁷ activating G-protein-mediated adenylyl cyclase that causes an intracellular elevation of cAMP.¹⁹¹ RXFP1 is expressed in granulosa cells and cumulus cells of antral follicles.¹⁹² It is assumed that relaxin is involved in oocyte maturation and influences granulosa/cumulus cell function in a paracrine fashion.¹⁹³

Insulin-like peptide 3

Insulin-like peptide 3 (INSL3) is structurally related to relaxin.¹⁹⁴ In the ovary, INSL3 is mainly produced in theca interna cells of antral follicles¹⁹⁵ and also in the corpus luteum.¹⁹⁶ INSL3 knock-out mice have reduced numbers of antral follicles, fewer corpora lutea and small litter sizes, pointing to that INSL3 plays an important role in promoting the number of growing antral follicles and ovulation.¹⁹⁷ In addition, INSL3 exerts an anti-apoptotic and pro-survival effect on growing antral follicles.¹⁹⁸ Glister et al., 2013¹⁹⁹ demonstrated that INSL3 is involved in a paracrine/autocrine feedback system that regulates androgen production in theca cells.

Antral follicles are the major source of circulating INSL3 in non-pregnant female mammals²⁰⁰ suggesting that INSL3 reflects the growth of antral follicles.²⁰¹ Accordingly, the number of growing antral follicles influences the level of circulating INSL3 and hence, circulating INSL3 levels are reduced in women with low ovarian reserve, while circulating INSL3 levels are significantly elevated in women with PCOS (Figure 11 to follow).²⁰¹

If there are any lessons to be learned at this time, it is that optimal gonadotropin hormonal action is highly contingent on the input of tissue-based regulatory principles. According to this view, gonadotropins may not be the omnipotent agents they were once thought to be. Rather, gonadotropins may best be viewed as 'team players' and as initiators of a cascade of events facilitated, attenuated, or mediated through interaction with putative intraovarian regulators. It is the special case of IGF-I that best illustrates the role of a tissue-based modulator in that optimal gonadotropin hormonal action is clearly highly dependent on the availability of IGF-I and the consequent amplification of gonadotropin hormonal action.¹⁵ In contrast, putative intraovarian regulators exemplified by TGF-α may attenuate gonadotropin hormonal differentiation in the interest of continued proliferative ability.

Another role for putative intraovarian regulators, exemplified by IL-1, is that of mediation of gonadotropin action. According to this view, IL-1 constitutes an extension of the gonadotropin signal, possibly one of several more distal effectors, the overall mission of which may well be the conveyance of the message (or portions thereof) imparted by the midcycle surge.

The development of the ovarian follicle is a continuum of growth and differentiation of at least three distinct cell types: thecal cells, granulosa cells, and oocytes. Much depends on the localization and timing of expression of the regulatory principles. Of equal importance is the ability of the target cell to receive and respond to the regulatory signal. Activin elicits a stimulatory and an inhibitory response, depending on the cell type being studied¹⁰⁴ and the developmental stage of the follicle.¹⁰² This duality of action may be explained if the activin receptor isoforms¹⁰¹ prove to have a specific cell-type and developmental-stage distribution. The action of a given regulatory factor, such as EGF/TGF, can also be influenced by the presence or absence of other factors.²⁰² The ability of IL-1²⁰³ and nerve growth factor²⁰⁴ to alter EGF/TGF binding in nonovarian cell types is an example of how one growth factor can impinge on the actions of another.

It is the net balance representing the integration of multiple transduction pathways (Fig. 12) and often opposing signals that determines final gonadotropin hormonal action. Moreover, a given intraovarian growth factor may play several roles, depending on its local concentration, availability of its receptors or binding proteins, the cell population with which it interacts, and the precise timing of that interaction. There is every reason to believe that future studies may reveal other modes of interaction between trophic ovarian principles and tissue-based regulatory elements. It is with a strong sense of excitement that future work in this evolving area is anticipated.

Fig. 12. Signal transduction pathways in the ovary

Survival, loss and activation of primordial follicles is mainly controlled by PI3K-signaling.^{205, 206} Oocyte specific deletion of Pten (phosphatase and tensin homolog deleted on chromosome ten), which is a negative regulator of PI3K or Pdk1²⁰⁷ results in a global activation of the entire primordial follicle pool²⁰⁸ leading to premature ovarian failure (POF) in Pten-deficient mice. Hence, a basal level of PI3K activation in oocytes is essential for the maintenance of the primordial follicle pool. In addition, AMH inhibits the PI3K induced activation of the primordial follicle pool and hence balances together with PI3K the ovarian reserve. Destruction of bigger, more mature follicles, for example through the administration of ovotoxic agents, diminishes AMH and thus, the loss of suppression leads through PI3K pathway activity to an accelerated activation of the primordial follicle pool, resulting in a “burn-out” effect of the ovarian reserve (Figure 13 to follow).²⁰⁹

Mammalian target of rapamycin complex 1 (mTORC1) in oocytes leads to premature activation of primordial follicles.²¹⁰ Treatment of mice with oocyte specific deletion of Pten with rapamycin, an mTORC1- inhibitor, retains a significant amount of the ovarian reserve in mice with excessive activation of the primordial follicle pool. Hence, the suppression of mTORC1 signaling pathway is important in the maintenance of the dormant primordial follicle pool (Figure 14 to follow).²¹⁰

In the ovary it was shown that FSH leads to accumulation of antihypoxia inducible factor-1 alpha (anti-HIF-1α) protein in granulosa cells and rapamycin, an mTOR inhibitor and PI3 kinase inhibitor, inhibits the FSH-stimulated HIF-1 activity.²¹¹ In addition, PI3 kinase/AKT-mediated activation of mTOR and phosphorylation of FOXO1 are essential for the FSH-stimulated HIF-1 induced up-regulation of VEGF-A.²¹² In the corpus luteum the expression of HIF-1α is highest in luteal granulosa cells during luteal formation, but absent in the fully functional corpus luteum.²¹³

During folliculogenesis an increased number of follicular blood vessels are regulated by different factors, most notably by vascular endothelial growth factor-A (VEGF-A).²¹⁴ VEGF-A is expressed in granulosa cells and theca cells of secondary follicles and is important for follicular maturation as the blocking of VEGF-A inhibits follicular growth.^{215, 216} After the LH-surge VEGF-A is up-regulated in luteal granulosa cells in the corpus luteum and remains expressed until the mid to late luteal phase.²¹⁷ Inhibition of VEGF-A disrupts ovulation, blocks vascularization of the corpus luteum and suppresses progesterone secretion from the corpus luteum.²¹⁸ On the other side of the spectrum, exaggerated levels of intraovarian VEGF play a cardinal pathophysiologic role due to increasing permeability of fluid through the ovarian blood vessels (therefore also called vascular permeability factor, VPF) the main modulator of the iatrogenic, and possibly serious, ovarian hyperstimulation syndrome (OHSS).

Fibroblast growth factor-2 (FGF-2) is expressed especially during the follicular-luteal transition in the ovary, but is not expressed in granulosa cells and theca cells in the ovary until the antral stages.²¹⁹ Inhibition of the FGF-receptor almost completely prevents the formation of the luteal endothelial networks.²²⁰ FGF-2 is expressed in oocytes of primordial and primary follicles²²¹ and promotes the transition from the primordial to the primary follicular stages, pre-antral follicular growth and recruitment of theca cells.²²² FGF-2 is expressed in theca interna cells and granulosa cells of antral follicles and modulates the action of VEGF-A.²²³ During the later stages of pre-ovulatory follicle development VEGF-A and FGF-2 expression increases.²²⁴ It has been shown that FGF-2 expression dramatically increases following the LH-surge and FGF-2 translocates from thecal endothelial cells to the nucleolus of granulosa cells.²²⁵ FGF-2 is critical for the formation of luteal endothelial networks.²²⁶ It is suggested that the increase in FGF-2 at the time during the follicular-luteal transition is important to stimulate the tissue remodeling after ovulation that accompanies angiogenesis.

Adipokines are biologically important molecules that are secreted from adipose tissue. Adipokines modulate glucose and lipid metabolism as well as insulin sensitivity. Adipokines are molecules such as leptin, adiponectin, resistin, chemerin, and apelin.²²⁷

Leptin

Leptin is suggested to be an important signal in female reproduction, including control of ovarian function, beside its role in the regulation of body weight and energy expenditure.²²⁸ In the ovary, leptin receptor Ob-R is expressed in both granulosa cells and theca cells. It has been shown that leptin counteracts the synergistic effect of IGF-1 on FSH-stimulated estradiol and progesterone production in granulosa cells.²²⁹ In addition, leptin receptor Ob-R is also expressed in the oocyte.²³⁰ Leptin increases the rate of meiotic resumption in pre-ovulatory follicle enclosed oocytes, probably via indirect actions on theca cells.²³¹

Adiponectin

Adiponectin activates peroxisome proliferation-activated receptors a (PPARa) and AMP-activated protein kinase. Adiponectin acts through two receptors, AdipoR1 and AdipoR2. Adiponectin and its receptors are expressed in the ovary.^{232, 233} Adiponectin decreases androgen and progesterone production indirect by insulin in theca cells, while in granulosa cells adiponectin increases progesterone and estradiol secretion in response to IGF-1.²³²

Resistin

Resistin activates 17-α-hydroxylase activity in theca cells in the presence of forskolin, suggesting a role of resistin in the regulation of androgen production in theca cells.²³⁴ Furthermore, in granulosa cells resistin modulates steroidogenesis and proliferation in response to IGF-1 and also in the basal state.²³⁵

Chemerin

Chemerin decreases IGF-1 induced progesterone and estradiol production through a decrease in the phosphorylation of IGF-1R b subunit and MAPK ERK1/2 signaling pathway.²³⁶

Apelin

Apelin is expressed in theca cells, while its receptor, AP J receptor, is expressed in both granulosa cells and theca cells.²³⁷ In granulosa cells, the increase in AP J receptor expression correlates with follicular atresia, while in theca cells apelin and AP J receptor are induced by LH.²³⁷

Neurotrophins and their receptors are expressed in the ovary²³⁸ and influence both ovarian somatic cells and oocytes.

Glial cell line derived neurotrophic factor

Glial cell line derived neurotrophic factor (GDNF), a distant member of the TGF-beta superfamily and known for promoting the survival and differentiation of peripheral and central neurons²³⁹ is expressed in the ovary.²⁴⁰ GDNF is expressed in ovarian granulosa and stromal cells as well as in oocytes.²⁴¹ GDNF stimulates oocyte nuclear and cytoplasmic maturation and cumulus cell expansion²⁴² and increases DAZL expression.²⁴³ However, it has also been suggested that GDNF may contribute to the onset of ovarian tumorigenesis in aging mutant mice carrying a disruption in the FSH receptor gene.²⁴¹

Brain derived neurotrophic factor

Brain derived neurotrophic factor (BDNF) is expressed in cumulus cells and in oocytes.²⁴⁴ BDNF promotes oocyte maturation.²⁴⁵ Furthermore, it has been shown that BDNF acts as a paracrine factor enhancing the extrusion of the first polar body.²⁴⁶

Nerve growth factor

Nerve growth factor (NGF) is expressed in human granulosa cells and oocytes.²⁴⁷ Injection of NGF into murine oocytes enhances the ability of oocytes to form parthenogenetic pronuclei in both oocytes within COCs and denuded oocytes.²⁴⁸