a. Functional changes in follicular cells

Only follicles that express a large number of LH receptors on granulosa cells' surface are able to ovulate in response to the pre-ovulatory LH peak (table 1). LH binds to its receptors and follicular cells respond to this signal by regulating the expression of several genes (figure 1). The binding of LH to its receptor increases intracellular levels of cyclic adenosine monophosphate (cAMP) and inositol tri-phosphate (Piquette et al., 1991), but this latter pathway is only a potentiating pathway (figure 1) and activated only in the presence of high hormone concentrations. Thus, protein kinase A (PKA) activated by cAMP, induces a cascade of phosphorylation and activation of transcription factors (CREB, SP1) that will inhibit or activate the expression of target genes coding for steroidogenesis enzymes. Thus, in the hours following circulating LH increase, serum estrogen and androgen levels collapse, while progesterone levels rise dramatically. In response to increased nitric oxide (NO; vasodilator activity) production, follicle theca becomes edematous and hyperhematous, increasing pre-ovulatory follicle blood flow and vascular permeability. Simultaneously, granulosa cells lose their FSH receptors, while LH receptor expression decreases transiently. In the rat, this "reprogramming" lasts approximately 7 hours (Richards and Hedin, 1988).

b. Follicular wall rupture

The ovulation process is associated with an inflammatory response type. Ovarian synthesis of inflammatory cytokines (IL, TNFα), prostaglandins and cortisol (with anti-inflammatory action) is enhanced in the pre-ovulatory follicle at the time of ovulation in mice, rabbits, sows, cows, mares and woman. Interleukin-1 (IL-1) is capable of inducing follicle rupture ex vivo, on infused ovaries of rat and rabbit (Brännström et al., 1993; Takehara et al., 1994) and in vivo in the mare after intrafollicular injection of dominant follicles (Martoriati et al., 2003).

The mechanism of action of cytokines in pre-ovulatory maturation is probably similar to that observed during inflammation. In particular, they stimulate the production of prostaglandins, NO, and the activity of proteolytic enzymes. In addition, they are able to modulate steroidogenesis and are also involved in the maturation of the oocyte-cumulus complex in rabbits and mares (Gérard et al., 2004).

Rupture of the follicular wall and expulsion of the oocyte into the oviduct requires the action of proteolytic enzymes that degrade the extracellular matrix. The local production of plasminogen activator, plasmin and collagenases (metalloproteinases MMP1 and MMP2) increases considerably. The plasminogen activator converts plasminogen into plasmin, which in turn activates collagenases (Beers et al., 1975). The pre-ovulatory increase in the intrafollicular concentration of progesterone stimulates the synthesis and activity of two other proteases ADAMTS1 and CTSL1, which are involved in proteolysis. In parallel, the activity of serine protease inhibitors (SERPIN) and matrix metalloprotease inhibitors (TIMP1 and TIMP2) increases. The expression of these metalloproteases and TIMPs is regulated by steroids and prostaglandins. Thus, the expression of proteases and anti-proteases in the pre-ovulatory follicle allows temporal and spatial regulation of proteolytic activity prior to rupture (Curry and Smith, 2006).

c. Maturation of the oocyte-cumulus complex

In response to LH pre-ovulatory peak, the oocyte, blocked at the prophase stage of the first meiotic division, enters the maturation phase by resuming its meiosis. This stage called GVBD (germinal vesicle break-down) is visualized by the rupture of the nuclear membrane. In most mammalian species, the oocyte evolves to the metaphase stage of the second meiotic division and remains blocked at this stage until fertilization. The molecular mechanisms involved in nuclear and cytoplasmic oocyte maturation are complex and will not be discussed in this review. LH induces cumulus expansion, a phenomenon necessary for meiotic maturation and the acquisition of competence in oocyte development. When the pre-ovulatory follicle ruptures, the follicular fluid is expelled taking down with it the oocyte-cumulus complex that will be captured by the oviduct.

a. Role of GnRH (Gonadotropin Releasing Hormone)

The synthesis and release of the two gonadotropic hormones LH and FSH are dependent on the hypothalamic neurohormone, GnRH (see chapter 2.3.a). Its role has been well established through the use of GnRH antagonists (Brooks et al., 1993) or anti-GnRH antibodies (Caraty et al., 1984; Molter-Gérard et al., 1999). These treatments induce immediate suppression of LH release, disappearance of the pre-ovulatory peak and blockade of ovulation. GnRH is released in a pulsatile manner by GnRH neurons and reaches the pituitary via the hypothalamic-pituitary portal blood system (figure 2). It in turn induces a pulsatile release of gonadotropic hormones. The frequency of GnRH pulses varies during the estrous cycle (figure 3): in ewes, the average frequency is of one pulse/hour during the follicular phase, and then becomes continuous in the period prior to ovulation. After ovulation, this frequency decreases and reach an average of one pulse every six hours during the luteal phase in the ewe (figure 3). The pulse frequency determines the pattern of gonadotropic hormones release. A high frequency favours the release of LH, whereas a low frequency favours the expression and release of FSH (figure 3) (Dalkin et al., 1989). During the pre-ovulatory peak of LH, there is a small amplitude FSH peak (figure 3) which allows the recruitment of follicles for the next cycle. While the synthesis and release of LH is intimately dependent on GnRH, FSH synthesis and relase depend also on other regulations (see §2.2.4).

b. GnRH acts via a specific receptor

GnRH acts in the pituitary via a specific receptor (GnRH-R) located on the surface of gonadotropic cells. GnRH-R belongs to the family of G protein-coupled receptors. The binding of GnRH to its receptor initiates an intracellular signalling cascade involving several proteins whose activation induces the transcription of genes coding for LH and FSH and/or the release of these hormones. The importance of GnRH-R is highlighted in cases of hypogonadism induced by mutations present on the GnRH-R gene inactivating its functionality (Chevrier et al., 2011). The number of GnRH-R receptors on the surface of gonadotropic cells varies with the frequency of GnRH pulses. High frequencies increase the number of receptors while low frequencies decrease the number of receptors (Katt et al., 1985). Different intracellular signalling pathways are thought to be activated as a function of the number of receptors, and thus the pulse frequency, leading to differential expression of the LHß and FSHß subunits and the release of LH and FSH (Stamatiades and Kaiser, 2018).

c. The essential role of estradiol

At the beginning of the follicular phase, the circulating concentrations of steroid hormones, mainly 17-β-estradiol (estradiol or E2), produced by the follicles remain low. E2 then exerts negative feedback control on the hypothalamic-pituitary complex by inhibiting the secretion of GnRH. At the end of the follicular phase, the concentration of E2 rises until it reaches a critical value at which the feedback becomes positive, generating the pre-ovulatory peak of LH, which causes ovulation.

d. Other factors regulating the secretion of gonadotropic hormones

In addition to the major role played by GnRH and estradiol, other factors are likely to modulate the synthesis and secretion of gonadotropic hormones. Growth factors of the Transforming Growth Factor ß (TGF-ß) family, in particular inhibin and activin, modulate the synthesis and release of FSH (figure 3). Inhibin B, produced by granulosa cells, inhibits the secretion of FSH at the end of the follicular phase. This action is essential for the selection of the dominant follicle(s). During the luteal phase, the corpus luteum secretes inhibin A, which reduces the release of FSH, thus preventing early maturation of the follicles. In addition, FSH synthesis is closely dependent on activin, produced in the pituitary gland. The absence of this factor leads to a drastic decrease in FSH concentrations whereas the massive injection of activin induces a shortening of the estrous cycle and an increase in the number of antrum follicles with premature superovulation (Erickson et al., 1995). The action of activin is modulated by a binding protein, follistatin, but also by inhibin, which prevents it from binding to its receptors. A slow frequency of GnRH pulses promotes the synthesis and release of activin (Dalkin et al., 1999).

a. The GnRH conductor of ovulation induction

GnRH is a ten-amino acid peptide neuro-hormone whose sequence is highly conserved in mammals (for a review see Duittoz and Prévot, 2014). This hormone has a particular horseshoe-shaped conformation that is important for binding to its receptor and for carrying out its biological activity. GnRH undergoes two post-translational modifications: cyclisation of N-terminal glutamine to pyro-glutamine and amidation of C-terminal glutamic acid (figure 4). These two post-translational modifications prevent the action of exopeptidases, but the peptide remains sensitive to endopeptidases capable of rapidly degrading it. As a consequence, GnRH half-life is less than five minutes and its secretion in small quantities makes its detection in peripheral blood impossible. GnRH neurons are distributed in a particular way within brain structures, they form a hypothalamic continuum with a diffuse distribution from the pre-optic area (PAO) to the Ventromedial Hypothalamus (VMH) (Wray, 2010). GnRH neurons project the majority of their axons into the median eminence (ME), in the vicinity of the capillaries of the hypothalamic-pituitary portal system (Duittoz and Prévot, 2014) (figure 5).

During oestrus, high concentrations of E2 induce a GnRH peak which leads to the pre-ovulatory peak of LH (Clarke et al., 1989; Moenter et al., 1991) but GnRH neurons do not express oestrogen receptor alpha involved in the positive retrocontrol, so this effect must be mediated by an intermediate system.

b. Role of kisspeptin in triggering spontaneous ovulation

The expression of KISS-1 gene results in a pre-pro-peptide, which is then cut into different peptides ranging from 54 amino acids for the longest, to 10 amino acids for the shortest. The set of these peptides is generally known as kisspeptin (Kp) (figure 6). The essential role of Kp in the control of reproduction was discovered through the study of mutations in the KISS1R gene encoding the GPR54 receptor, inducing a loss of function associated with hypogonadotropic hypogonadism in infertile patients (de Roux et al., 2003). Administration of Kp in ewes results in the secretion of GnRH and the gonadotropins LH and FSH (Caraty et al., 2007). Kp acts directly on GnRH neurons, the majority of which express the GPR54 receptor and produce an increase in their electrical activity (Messager et al., 2005). Kp is secreted by two distinct populations of neurons located one in the arcuate nucleus (ArcN) and the other in the POA in ewes or the periventricular anteroventricular area (AVPV) in rodents. Kp neurons in POA/AVPV show a marked sexual dimorphism. In males, there are very few cell bodies in POA/AVPV while in females this region is rich in Kp expressing neurons. The somas of GnRH neurons are mostly contacted by POA Kp while GnRH neurons fibres located in ME are mostly contacted by ArcN Kp neurons (Caraty and de Roux, 2014). Both Kp neurons populations express the α receptor for estrogen, and thus could support the transmission of positive E2 feedback (Franceschini et al., 2006; Smith, 2008).

c. Influence of photoperiod

One of the environmental signals exerting a major control on the secretion of GnRH is the photoperiod. Photoperiod refers to variations in day length over the course of a year. It underlies the alternation of sexual activity and sexual rest phases in so-called seasonal species (e.g. sheep, goat, horse). This temporal marker allows these species to anticipate the period of the year that is favourable for the development of offspring.

The light message associated with the photoperiod is picked up by the retina and transmitted to the pineal gland through a multisynaptic pathway. In the absence of light, the pineal gland releases melatonin, this secretion is inhibited in the presence of light through the neural pathway from the retina. Melatonin does not act directly on GnRH neurons since they do not possess the type 1 melatonin receptor (MT1 involved in the control of reproductive activity on long days) (Migaud et al., 2016). The precise mechanism of transmission of the melatonin message to GnRH neurons is not yet clearly established. Nevertheless, some crucial players seem to emerge. In particular, melatonin modulates the local concentrations of thyroid hormones via control of deiodinases, enzymes that convert thyroid hormone thyroxine (T4). Diodinases II (Dio II) and III (Dio III) are respectively responsible for the conversion of T4 into its active form triiodo thyronine (T3) or into inactive forms diiodo thyronine (T2) and reverse T3 (rT3) (Dardente et al., 2014). However, the expressions of Dio II and Dio III vary according to the time of year. In ewes, Dio II is expressed more during the breeding season and Dio III during the sexual resting season. These variations in deiodinase expression are themselves dependent on variations in Thyroid stimulating hormone (TSH) expressed locally in the pars tuberalis of the pituitary gland. TSH secretion is known to be directly dependent on melatonin, whose MT1 receptor is strongly expressed in the pars tuberalis. However, two major questions remain unanswered. The first is what is the true physiological importance of this mechanism? Indeed, TSH/deiodinase/T3 pathway induced by long days is present not only in animals that breed on long days (e.g. hamster, quail...) but also in animals that breed on short days (e.g. sheep) and in non-seasonal animals (e.g. mice). The second question is how do local variations of T3 in the mediobasal hypothalamus control GnRH secretion? Several studies underline the role of populations of neurons expressing RF-amide type peptides: the arcuate nucleus with Kp neurons and the ventromedian and dorsomedian hypothalamus where neurons expressing RFRP-3 (also known as gonadotropin-inhibitory hormone, GnIH) are found (Dardente et al., 2014). Kp neurons may thus represent a link between changes in thyroid hormone levels and reproductive axis function. T3 would act by blocking the estrogen/Kp dialogue essential for the secretion of GnRH (Dardente, 2011; figure 7).

d. Importance of the glial environment

GnRH neurons are closely associated with tanycytes, a particular type of glial cell that lines the third ventricle and the median eminence. Tanycytes undergo structural changes during the ovarian cycle, they retract their extensions or feet, allowing direct contact of GnRH neurons axonal endings onto blood vessels of the hypothalamic-pituitary portal system, which would facilitate the diffusion of GnRH into the blood (Prévot et al., 1999). These changes in tanycyte morphology depend on the production of nitric oxide (NO) by endothelial cells (de Seranno et al., 2010). E2 leads to the retraction of tanycyte extensions notably by stimulating the expression of eNOS (endothelial nitrite oxide synthase) in endothelial cells. In addition, E2 also acts directly on tanycytes by stimulating the expression of COX-1 and COX-2 cyclooxygenases, stimulating the production and secretion of prostaglandin E2 (PGE2) which in turn, stimulates GnRH neurons electrical activity (Clasadonte et al., 2011). Other factors such as neuregulins and Transforming Growth Factor TGF- via their erbB1 and erbB2 receptors present on tanycytes, stimulate the expression of COX and the production of PGE2 (Prévot et al., 2010).

Glial cells forming the microenvironment of GnRH neurons are coupled together via gap junctions. Blocking communication through gap junctions causes a significant decrease in pulsatile GnRH secretion suggesting that in the basal state, communication via gap junctions between glial cells is essential for adequate GnRH secretion. (Pinet-Charvet et al., 2016).

e. Role of the Beta Nerve Growth Factor in Ovulation

In mammals with induced ovulation such as llama, alpaca and dromedary, ovulation is triggered by a factor present in seminal plasma: OIF (see for review Adams et al., 2016). In alpacas, administration of OIF (1mg equivalent β-NGF) by intramuscular route causes ovulation by a mechanism involving the secretion of GnRH (Silva et al., 2011). Other studies suggest that OIF is also present in spontaneously ovulating species. Indeed, intramuscular injection of seminal plasma from bulls, stallions and boars to female llama, provoke ovulation in 26%, 29% and 18% of cases respectively, suggesting that the seminal plasma of these species also contains molecules with OIF activity (Bogle et al., 2011).

Proteomic analysis of alpaca seminal plasma allowed the identification of the chemical nature of OIF which corresponds to the "β-Nerve Growth Factor" (β-NGF) (Kershaw-Young et al., 2012, Ratto et al., 2012). β-NGF is a neurotrophin identified by Rita Levi-Montalcini (Nobel Prize in Physiology in 1986). β-NGF has two types of receptors, TrkA and P75NTR. β-NGF plays a role in the survival, proliferation and differentiation of neurons (for review Levi-Montalcini et al. , 1996).

Like Kp, β-NGF could play its role in ovulation induction by stimulating the secretion of GnRH. In support of this hypothesis, in vitro studies carried out on embryonic mouse GnRH neurons primary induces an increase in GnRH neurons electrical activity (Pinet-Charvet and Duittoz, 2015).

f. Role of leptin

Metabolic status also play a role in the regulation of GnRH secretion and in the induction of ovulation. Numerous studies show that fertility is closely linked to body condition and nutritional status (Dupont et al., 2014). The link between weight and body condition and the onset of puberty in heifers or postpartum fertility in cows has been known for more than 50 years, for review (D'Occhio et al., 2019). Numerous metabolic factors have been shown to influence either directly the ovarian follicles or indirectly via regulation of the hypothalamo-hypophysio-gonadal axis. Two main actors are hormonal in nature and have been discovered over the last 20 years, leptin, which is secreted by fat tissue, and ghrelin, which is secreted by the stomach.

Information on the state of an individual's lipid reserves is mediated primarily by leptin. This hormone is known to act on gonads and pituitary cells. It also seems to play a role in the secretion of GnRH, but indirectly. Since GnRH neurons do not express the leptin receptor, it could act via Kp neurons. Indeed, Kp neurons express leptin receptors (Backholer et al., 2010; Hausman et al., 2012). Other neurons may play this role, in sheep and rodents, ArcN Neuropeptide Y (NPY) neurons make synaptic contacts on GnRH neurons cell bodies (Norgren and Lehman, 1989; Tillet et al., 1989). NPY may have either stimulating or inhibitory effects on the activity of GnRH neurons via the Y4 and Y1 receptors, respectively (Roa and Herbison, 2012). These NPY neurons express leptin receptors and are involved in the control of energy balance (Baskin et al., 1999) (figure 7). Ghrelin is secreted by the stomach and acts on the hypothalamus; its secretion induces hunger feelings. Ghrelin acts directly on GnRH neurons by reducing their activity (Farkas et al., 2013).

a. Current methods

Two main methods can be distinguished according to the target: ovary and hypothalamic-pituitary complex (figure 8).

b. Action at the ovarian level

Physiologically, LH causes ovulation by promoting the rupture of the follicle wall through the cellular and molecular mechanisms described above. However, LH has a short plasma half-life (less than two hours, except in mares). LH is therefore not used in practice. However, in the human and equine species, there are placental hormones with LH activity, hCG ("human chorionic gonadotropin") and eCG ("equine chorionic Gonadotropin"). hCG is produced by trophoblastic cells, its post-translational modifications give it a long half-life between 24 and 48 hours. Thus, hCG has been granted marketing authorization for the induction of ovulation in domestic mammals (equine, bovine, ovine, caprine, porcine, dog and cat) under the name Chorulon®. However, hCG aminoacids sequence is different from that of LH from other mammalian species and therefore, antibodies will be produced by the immune system and will block hCG activity. Furthermore, hCG can induce ovarian hyperstimulation. eCG has LH activity in the mare but has FSH-like activity in other species. It is used under its former name, PMSG ("pregnant mare serum gonadotropin", e.g. Chrono Gest®) to stimulate follicular growth after synchronisation by progestagens (fluorogestone e.g. Synchropart®). Ovulation occurs spontaneously following the development of pre-ovulatory follicles. This method is used for natural breeding or AI in fresh semen mainly in small ruminants. The use of eCG poses the same problem as that of hCG, i.e. repeated injections favour the appearance of antibodies and the ineffectiveness of subsequent stimulation. In addition, the method of eCG production which involves bleeding of pregnant mares that are aborted, raises ethical issues that have recently been echoed in European media. hCG is purified from urine of pregnant women and does not pose any particular ethical problems.

c. Action at the hypothalamic-pituitary level

- GnRH Agonists

GnRH agonists will act on GnRH receptors expressed by gonadotropic cells and induce the secretion of LH and FSH. This stimulation can lead to a very important secretion of LH because the stocks of LH in the pituitary gland are much higher than the quantity of LH mobilized during a physiological stimulation by endogenous GnRH. GnRH agonists are peptides derivate from GnRH to prolong half-life. Of these compounds, only two have veterinary market authorisation in France, deslorelin (Ovuplant®) and buserelin (Receptal®). As the chemical structure of these agonists is based on the amino acid chain of GnRH, the probability that they induce an immune reaction is very low and to date, no such effect has been reported in domestic mamals. These compounds should be used with caution, because if their persitence is too long, the pituitary gland will desensitize and will no longer respond to endogenous GnRH. If, following induction of ovulation by means of a deslorelin implant, insemination is not followed by fertilisation, the next cycle will be delayed. Deslorelin in the form of an implant is used to induce an estrus state in heifers in order to synchronize them (Padula, 2005), but also in dogs and stallions to obtain chemical castration. In the human species, goserelin, triptorelin and leucoprorelin, three GnRH agonists are used orally or by injection to induce chemical castration and to lower circulating estrogen levels in the adjunctive treatment of estrogen-dependent tumours. Thus, the use of GnRH agonists to induce ovulation can be tricky and lead to contradictory.

- Estrogen

Administration of a high dose of E2 induces ovulation. The mechanism involves both hypothalamic and pituitary effects. High-dose E2 exerts positive feedback control on the hypothalamus and results in massive secretion of GnRH. It exerts positive feedback control on the pituitary gland by potentiating the secretion of LH. Unfortunately, the use of E2 or analogues poses many problems for the environment. Indeed, the analogues used are not very degradable in the organism and are eliminated in active form. They accumulate in surface waters and are not necessarily disposed of in water treatment plants. Estrogenic contamination of watercourses is a major environmental problem. The use of oestrogens for the control of reproduction has been banned in Europe since 2003 (Directive 2003/74/EC of the European Parliament). Nevertheless, their use outside Europe is developing, particularly in protocols without heat detection and programmed AI (Torres-Júnior et al., 2014).

a. Stimulation of GnRH secretion by Kp

In spontaneously ovulating species, ovulation is triggered by stimulation of GnRH neurons by Kp. In human in vitro fertilization protocols Kp54 has been successfully used to induce oocyte maturation when using a standard superovulation protocol (Jayasena et al., 2014). Farm animal studies have so far focused on ewes with promising results. During anestrus season, Kp10 intra venous infusion lasting 24 hours or more is capable of reactivating the hypothalomo-hypophysis-gonad axis and triggering ovulation (Caraty et al., 2007). More recently we have developed a Kp10 analogue, called C6, with an improved pharmacological profile (Beltramo et al., 2015). A single intramuscular injection of C6 is capable, following pre-treatment with a progesterone analogue, of inducing pre-ovulatory peaks of LH and fertile ovulation in ewes (Decourt et al., 2016a). These results confirm that it is possible to control reproduction by acting directly on GnRH neurons with Kp analogues.

b. Stimulation of GnRH secretion by β-NGF

In camelids, species with induced ovulation, β-NGF is the trigger for ovulation. However, it seems that the β-NGF is also capable of inducing ovulation in spontaneously ovulating mammals. Thus, administration of alpaca seminal plasma to prepubertal mice in an ovarian stimulation protocol can induce ovulation (Bogle et al., 2011). β-NGF increases LH secretion and improves corpus luteum function in cattle (Tribulo et al. , 2015). In another species with induced ovulation, the rabbit, recent work shows the coexistence of two mechanisms (Maranesi et al., 2018). The first involves the β-NGF of the seminal plasma and the second involves the production of β-NGF by the uterus, the latter acting on the visceral sensory fibers that express the TrkA receptor, and sending a signal, as yet unidentified, to the hypothalamus. Although the mechanisms involved in the effects of β-NGF on ovulation have not yet been clarified, the results show the interest at deciphering the molecular mechanisms involved. Furthermore, an important point for further exploration of the potential of β-NGF to control reproduction in domestic mammals will be the creation of small synthetic agonists that will make the use of this strategy economically feasible. The first steps in this direction have been taken for example with the synthesis of small molecules such as LM11A-31 (Massa et al., 2006) and BNN27 (Pediaditakis et al., 2016).