Neuroanatomical Framework of the Metabolic Control of Reproduction

Abstract

A minimum amount of energy is required for basic physiological processes, such as protein biosynthesis, thermoregulation, locomotion, cardiovascular function, and digestion. However, for reproductive function and survival of the species, extra energy stores are necessary. Production of sex hormones and gametes, pubertal development, pregnancy, lactation, and parental care all require energy reserves. Thus the physiological systems that control energy homeostasis and reproductive function coevolved in mammals to support both individual health and species subsistence. In this review, we aim to gather scientific knowledge produced by laboratories around the world on the role of the brain in integrating metabolism and reproduction. We describe essential neuronal networks, highlighting key nodes and potential downstream targets. Novel animal models and genetic tools have produced substantial advances, but critical gaps remain. In times of soaring worldwide obesity and metabolic dysfunction, understanding the mechanisms by which metabolic stress alters reproductive physiology has become crucial for human health.

I. INTRODUCTION

Animal life requires sufficient energy for basic physiological processes (e.g., protein biosynthesis, thermoregulation, locomotion, cardiovascular and brain function, digestion). However, reproducing requires additional energy use. The production of sex hormones and gametes, pubertal development, searching and competing for sex mates, territoriality, sexual behavior, pregnancy, lactation, and parental care all demand extra energy stores. For this reason, the physiological systems that control energy homeostasis and reproductive function coevolved to support successful reproduction and the survival of the species (138, 144, 378, 442).

Different species have developed distinct reproductive strategies to adapt to specific environmental constraints (83, 140, 399). These strategies synchronize reproduction with the best environmental conditions for bearing and feeding offspring. While metabolic signals are crucial for fertility in many species, for others, photoperiods, temperature, and humidity play stronger roles. In this review, we will not discuss these alternative environmental limitations. Instead, we will emphasize the scientific literature focusing on the brain circuitry that coordinates energy availability with reproduction and the mechanisms by which metabolic stress may alter fertility.

The modulatory effect of nutritional status on reproductive physiology has been known for centuries, from Aristotle to Darwin and among farmers. This idea originated from observations that domesticated animals with regular access to food show early puberty and have improved reproductive health. In seminal experimental studies, Kennedy and Mitra (240) found that the time of puberty onset in rats correlates with body size, not chronological age. Subsequently, Frisch (157, 158) proposed that a critical threshold of body fat is necessary for sexual maturation, based on epidemiological studies in humans.

In the past decade, the use of murine models and genetic tools has rapidly expanded our understanding of the key mechanisms associated with energy homeostasis and reproduction. The brain plays a critical role in sensing the internal environment, integrating these complex systems, and producing the appropriate neuroendocrine and behavioral responses. Following a comprehensive analysis of findings from laboratories around the world, we will propose a framework for the brain circuitry engaged in the metabolic control of reproductive physiology. Despite substantial advances, many questions remain unanswered. By identifying these gaps, we hope to inform future research. Finally, we will highlight the relevance of the scientific data collected in animal models to human reproductive health.

II. NEURAL SUBSTRATES OF REPRODUCTIVE PHYSIOLOGY

The neural control of reproductive physiology depends on sensory signals from the internal and external environment, brain integrative centers, and motor outputs. Sensory signals include sex steroids, metabolites, and peptide hormones that convey information about the individual’s reproductive status (e.g., age, estrous/menstrual stage, pregnancy) and environmental cues (e.g., presence of sex mates, nutrition, and photoperiod in seasonal breeders) to specific brain nuclei housing first-order neurons. These physiological signals are then processed in integrative sites that orchestrate a series of coordinated motor responses via the neuroendocrine system and behavioral output.

A. Neuroendocrine Control

The ultimate goal of sexual reproduction is gamete contact (oocytes and spermatocytes). In viviparous females, oogenesis starts during embryonic development, is arrested around birth, and resumes at puberty; in males, spermatogenesis is initiated during pubertal development (363, 420, 441). Puberty is defined as the transition to sexual maturity, conferring the ability to reproduce. It is a complex physiological process triggered by the activation of the hypothalamo-pituitary-gonadal (HPG) axis (338, 363, 420). The gonadotropin releasing hormone (GnRH) neurons are the final neural path for activation and coordinated control of the neuroendocrine axis during pubertal maturation and in the adult life. Numbering ~800 in rodents and 2,000 in humans, GnRH neurons are widespread in the hypothalamus of most species and in the medial septum of rodents (244, 414, 416).

GnRH neurons originate in the olfactory epithelium and migrate into the brain during embryonic development (387, 404, 446, 499). Migrating GnRH neurons follow the vomeronasal nerves caudally into the hypothalamus (and medial septum, in rodents), extending processes into the vascular organ of the lamina terminalis and the median eminence (446, 506). Before birth, GnRH cell bodies reach their final destination (the hypothalamus and medial septum), and GnRH fibers densely innervate the external layer of the median eminence, ending in the portal vasculature.

GnRH neurons exhibit several prenatal and early postnatal periods of activity, and then become relatively quiescent (361) until pubertal maturation, when the “reactivation” of GnRH pulsatile secretion triggers puberty in both rodents and humans (338, 363, 420). The influence of steroid hormones during the period of the central reactivation of GnRH release differs between species. Gonadal hormones are not required for GnRH pulse generator “reactivation” in primates, since puberty can occur despite gonadal dysgenesis or castration (95, 362, 365). In contrast, in rodents, gonadotropin secretion is suppressed by small amounts of gonadal steroids after birth through the juvenile period. In peripubertal mice and rats, low levels of steroids are no longer inhibitory (15, 16, 338, 458). The increase in GnRH pulsatile release induces gonadotropin [luteinizing hormone (LH) and follicle stimulating hormone (FSH)] synthesis and secretion from the gonadotropes in the anterior pituitary gland (adeno-hypophysis). Higher levels of gonadotropins stimulate the full development of the gonads, the synthesis and secretion of gonadal steroids, and the maturation of the gametes. Thus the precise and coordinated stimulation of GnRH neuronal activity and GnRH secretion triggers sexual maturation. However, the signals or physiological changes that increase GnRH pulsatile release at the onset of puberty are not completely known. For example, whether prepubertal GnRH neurons are under a high inhibitory restraint or low excitatory drive is still debated (86, 338). Several studies have shown that an increase in excitatory input and/or in the sensitivity of GnRH neurons to neurotransmitters (e.g., GABA, glutamate) or neuropeptides (e.g., kisspeptin) is crucial for puberty onset (51, 191, 338). Additionally, decreased sensitivity to the negative-feedback actions of estradiol and the removal of inhibitory forces also play a crucial role (86, 299, 338, 363, 441). Whether changes in neuronal inputs, receptor sensitivity, epigenomic or neuronal/glial remodeling are the drivers of these physiological changes and, if so, what triggers them is unknown. A variety of internal and external signals act as permissive factors for pubertal maturation (338, 368). Among them, nutrition and energy sufficiency are critical factors in this process (138, 144, 378).

Electron microscopy studies have shown that GnRH occupies dense core vesicles in most cellular compartments, including in cell bodies, dendrites, axons, and terminals (179, 270, 367). Therefore, dendrites and axons may release GnRH not only in the median eminence but also in the preoptic area, medial septum, and areas where neuronal processes reside (159, 178). Because of this combined input/output function of GnRH processes (dendrites and axons), investigators in the GnRH field have dubbed them “dendrons” (203). Neuromodulatory peptides such as kisspeptin and gonadotropin inhibitory hormone (GnIH) differentially regulate the release of GnRH in the median eminence and preoptic area (178, 255). GnRH may therefore be released in response to inputs arriving at the soma or directly at the terminals in the median eminence. Whether these differential effects are components of the same physiological output is still unknown.

GnRH is released into the median eminence in distinct modes depending on reproductive stage, such as puberty, phase of the estrous/menstrual cycle, and aging. Experiments using immortalized GnRH neurons and primary cell culture have shown that GnRH neurons maintain intrinsic activity sufficient for the release of GnRH in a pulsatile manner (257, 305, 486). Burst firing, synchronized calcium oscillations, and GnRH release continue to occur in GnRH neurons isolated from the embryonic olfactory placode of most mammals (131, 160, 443). These observations indicate that GnRH neurons have an intrinsic pulsatile activity and suggest that the coordination of the GnRH network controls gonadotropin secretion. In fact, neuroanatomical studies have shown that cell-to-cell contact between GnRH neurons creates an interconnected network that coordinates the activity of GnRH neurons as a whole (64, 65). Although these in vitro and anatomical findings are solid, several pieces of the puzzle are still missing. For example, it is unclear whether GnRH neurons express the GnRH receptor (482). In rhesus monkeys, the intracerebroventricular administration of GnRH or an antagonist (antide) failed to change multi-unit activity at the mediobasal hypothalamus that was synchronized with LH pulses, despite having an effect on LH secretion (341). In rats, whereas central GnRH induces LH secretion, it has no effect on LH pulsatility (44). Furthermore, the role of other neurotransmitters or neuromodulators coexpressed by GnRH cells in the synchronization of GnRH activity and release has not been demonstrated.

In most rodents and primates, changing levels of sex steroids modulate the activity of GnRH neurons across the estrous/menstrual cycle. In the initial phase of follicular growth, estradiol is low, exerting a negative feedback that maintains low GnRH and LH levels. With the maturation of the follicles, increased secretion of estradiol stimulates GnRH and LH secretion (positive feedback) (183, 202, 274, 311, 312). In concert with the actions of estradiol on the pituitary gland, this abrupt outpouring of GnRH triggers the preovulatory gonadotropin surge, ultimately leading to ovulation (183, 202, 274, 311, 312). The LH surge in rodents is triggered by the concurrence of a circadian signal with high levels of estradiol observed in the late follicular phase (79–81, 267) While progesterone has been implicated in increasing the size of the LH surge, it is not required in rodents; instead, it coordinates mating and ovulation (46, 53, 184, 185, 306). FIGURE 1 illustrates the main components of the HPG axis in humans.

FIGURE 1.

Schematic illustration showing the basic organization of the hypothalamo-pituitary-gonadal (HPG) axis of humans. Gonadotropin releasing hormone (GnRH) neurons located in the hypothalamus project to the median eminence where GnRH is released from the terminals into portal vessels. GnRH acts in gonadotropes located in the anterior pituitary gland inducing synthesis and secretion of gonadotropins, i.e., luteinizing hormone (LH) and follicle-stimulating hormone (FSH). The gonadotropins induce production of sex steroids that feed back into the pituitary gland and hypothalamus. -, Negative feedback; +, positive feedback in females.

The negative- and positive-feedback actions of estradiol on GnRH pulsatile release are achieved mainly via binding to estrogen receptor α (ERα), encoded by the Ers1 gene (99). Because Ers1 expression is undetectable in GnRH neurons, interneurons are essential for controlling the female reproductive cycle (333, 492). Kisspeptin is the most potent GnRH secretagogue identified in mammals. Several excellent reviews of the role of kisspeptin in the neuroendocrine control of reproductive function have been published in recent years and should be consulted for details (93, 269, 335, 359). Loss-of-function mutations in KISS1/Kiss1 or kisspeptin receptor (GPR54/Gpr54, also known as Kiss1r) genes cause infertility in humans and mice due to a lack of pubertal development and hypogonadotropic hypogonadism (107, 119, 262, 408, 450). A steady increase in hypothalamic Kiss1 and Gpr54 gene expression is observed across pubertal development, and administration of kisspeptin to juvenile rodents advances puberty, increases LH secretion, and induces ovulation (328, 411, 440). Kisspeptin acts directly on GnRH neurons to control their activity and neuropeptide release via two distinct neuronal populations: 1) kisspeptin neurons in the anteroventral periventricular and preoptic periventricular nuclei (AVPV/PVpo, also called the rostral periventricular area of the third ventricle, or RP3V); and 2) kisspeptin neurons in the arcuate nucleus (ARH) (202, 425, 426). Both populations coexpress Esr1 and are tightly regulated by estradiol (155, 170, 269, 356, 424). Generally, AVPV/PVpo neurons are believed to convey the positive-feedback actions of estradiol, while ARH neurons convey the negative-feedback actions of estradiol to GnRH neurons in rodents. AVPV/PVpo kisspeptin neurons predominate in females, project to GnRH neurons, have ion channels and Kiss1 expression modulated by estradiol levels throughout the estrous cycle, and are activated at the time of the surge (1, 85, 357, 427, 473).

Kisspeptin neurons in the ARH coexpress neurokinin B (NKB) along with its receptor NK3R (116, 329). NKB appears to serve as an autoregulatory signal for ARH kisspeptin neurons, stimulating kisspeptin output to GnRH neurons (373, 374). Humans with inactivated mutations of the genes encoding NKB (TAC3, whose ortholog is Tac2 in rodents) or its receptor, NK3R (TACR3), suffer a lack of pubertal maturation and central hypogonadism (449), a phenotype reminiscent of patients with mutations of KISS1 or KISS1R (GPR54) genes. ARH kisspeptin neurons also express dynorphin (57, 269), an endogenous opioid peptide that selectively binds the κ-opioid receptor (KOR). The administration of a KOR antagonist advances pubertal development and increases LH pulse frequency in juvenile and adult rats (316, 326). Because of the coexpression of these three neuropeptides, ARH kisspeptin neurons have been named KNDy neurons. Double immunoelectron microscopy has revealed that NKB, dynorphin, and kisspeptin are contained in individual neurosecretory vesicles within kisspeptin neurons, suggesting that the synthesis of individual KNDy peptides is differentially regulated (321). It has been proposed that, when synthesized, NKB primarily contributes to the onset of kisspeptin activation and, subsequently, the preferential production of dynorphin terminates that action in KNDy neurons.

Similar to estradiol, metabolic cues affect GnRH function primarily by acting on afferent neurons. This evolutionary choice increases complexity and redundancy, allowing a more integrated and precise control of motor/endocrine responses and support in case of disruption of one path. In this review, we will describe the neural circuitry engaged in transmitting to GnRH neurons information on the immediate nutritional status and the amount of available energy stores. We will also discuss the role of kisspeptin as one of the best-known conveyors of metabolic status to the HPG axis.

B. Behavioral Output

The production of sex steroids induced by the neuroendocrine reproductive axis allows or facilitates behavioral responses in both sexes. Reproductive behavior encompasses numerous processes including courtship, sexual, aggression, and parental behaviors, all of which are vital to ensuring the survival of the species.

Research using rodents as a model species has yielded quantifiable parameters of sexual behavior in males and females. Once the opposite sex is introduced, the cycle of copulatory activity begins with an arousal phase, during which the sex and physiological competence of the new animal is assessed through anogenital sniffing and increased locomotor activity (354). This chemosensory input is essential for initiating copulatory behavior (120). Sensory neurons in the olfactory epithelium sense volatile odorants and send projections to the main olfactory bulbs. Damage to this system eliminates copulation in mice, independent of sexual experience (238, 289, 390). In addition, sensory neurons in the vomeronasal organ along the ventral midline of the nasal cavity perceive nonvolatile pheromones during physical contact. Evidence shows that the accessory olfactory system motivates males to seek female urinary pheromones and augments their sexual arousal to promote intromission (258). Females also require continued accessory olfactory signaling for maximum sexual responsiveness (300).

Both estrogens and progesterone facilitate female mating behavior in many species including rodents and primates (114, 428, 471, 511). In rodents, female sexual arousal and receptivity correlate with a rise in estradiol and progesterone before ovulation. Estrogens (via aromatization of testosterone) are also the main regulators of male sexual behavior, although the degree to which they impact behavior is species-dependent (217, 388).

Main and accessory olfactory systems convey odorant signals (arousal cues) to sex steroid-primed brain circuits to the medial subdivision of the bed nucleus of the stria terminalis (BST), the medial nucleus of the amygdala (MEA), that in turn target the hypothalamic medial preoptic nucleus (MPN) and the ventral premammillary nucleus (PMV) (35, 96, 98, 126, 272, 396, 495). These areas are highly interconnected and converge into the ventrolateral subdivision of the hypothalamic ventromedial nucleus (VMHvl) that is essential for mating behavior in both sexes (32, 47, 48, 264) (FIGURE 2). Stimulation of the VMHvl facilitates lordosis and receptivity in females (253, 352), and VMHvl lesions abolish the lordotic response in females and the ability to mount in males (351). The VMHvl projects to the lateral and ventrolateral columns of the periaqueductal gray matter, hindbrain areas that integrate sensorial, endocrine, and autonomic inputs and modulate the motor responses of sexual behavior (66, 418).

FIGURE 2.

Schematic illustration showing coronal sections of the brain sites associated with the basic behavioral output in rodents. The presence of a potential sex mate is conveyed by odorants via the main and accessory olfactory systems to the medial subdivision of the bed nucleus of the stria terminalis (BST) and to the medial nucleus of the amygdala (MEA). These nuclei innervate hypothalamic sites associated with reproductive control including the medial preoptic nucleus (MPN) and the ventral premammillary nucleus (PMV). These areas share reciprocal connections and innervate (blue) the ventrolateral subdivision of the ventromedial nucleus of the hypothalamus (VMHvl), the main output (red) to midbrain and brainstem sites and behavioral control. Kisspeptin neurons (green) in the anteroventral periventricular nucleus (AVPV) also target the VMHvl conveying information about the reproductive status. ac, anterior commissure; ARH, arcuate nucleus; DMH, dorsomedial nucleus; ERα, estrogen receptor α; f, fornix; LV, lateral ventricle; MPO, medial preoptic area; MS, medial septum; PR, progesterone receptor.

The activation of the neuronal circuitry responsible for copulatory behaviors depends on the normal functioning of the HPG axis. Notably, AVPV kisspeptin projections to VMHvl are crucial for the function of this circuitry (198). Without stimulation of these circuits by sex steroids, copulation fails. Therefore, the behavioral aspects of fertility depend on the normal functioning of the HPG axis, and the factors that alter GnRH release are likely to modify sexual behavior. These factors include nutrition and energy stored as fat.

III. CNS INTEGRATION OF METABOLIC AND REPRODUCTIVE FUNCTION

A. Sensing Metabolic Cues

To integrate metabolism and reproduction, the central nervous system (CNS) relies primarily on inputs carrying information on the immediate nutritional state and the status of energy stores. These inputs, generally called “metabolic cues,” originate either from the breakdown of nutrients into metabolites and oxidizable fuels acting directly on the CNS (e.g., amino acids, glucose, free fatty acids) or from hormones secreted into the circulation by metabolic tissues. Among them, white adipocytes (e.g., leptin, adiponectin, resistin), the pancreatic islets (e.g., insulin, glucagon), and the digestive tract and accessory organs [e.g., ghrelin, glucagon-like peptide 1 (GLP1), fibroblast growth factor 21 (FGF21), insulin-like growth factor 1 (IGF1)] are the main sources of metabolically relevant hormones (FIGURE 3). In most cases, these hormones directly target first-order neurons in hypothalamic and brainstem nuclei upstream of GnRH neurons.

FIGURE 3.

Schematic illustration showing the metabolic cues and associated organs with function in the control of reproduction via actions in the central neurons system. Leptin (produced by white adipocytes), insulin (secreted by pancreatic islets), ghrelin and glucagon-like peptide 1 (GLP1, produced by the gastrointestinal tract), fibroblast growth factor 21 (FGF21) and insulin-like growth factor 1 (IGF1), both produced by the liver, and circulating metabolites [glucose, amino acids (AA), and free fatty acids (FFA)] target neurons in the hypothalamus or caudal medulla (glucose).

To exert their effects, the metabolic hormones must cross the blood-brain barrier (BBB) and access cognate receptors in the brain parenchyma. The BBB is a dynamic structure that regulates the influx and efflux of ions, nutrients, and hormones and restricts the access of pathogens (26, 236). It consists of distinct cells from the brain parenchyma (i.e., neurons, astrocytes, pericytes, glia) and specialized endothelial cells, together named the “neurovascular unit.” The BBB exists throughout the CNS with the exception of the circumventricular organs. The circumventricular organs contain fenestrated blood vessels that allow the diffusion and interchange of large molecules (i.e., peptides and hormones) between the neural tissue and the circulation, possibly without the need for active transport (52, 165, 228). Among the well-described circumventricular organs, the vascular organ of the lamina terminalis and the median eminence are of particular interest given their proximity to and functional relationship with reproductive regulatory neurons in the MPO, AVPV/PVpo, and ARH (202, 335, 359). At these sites, metabolic cues have access to the adjacent neural tissue and bind to their receptors. Outside the circumventricular organs, metabolic cues cross the BBB via two known mechanisms: 1) lipid-mediated free diffusion or 2) carrier- or receptor-mediated active transport. Most metabolic hormones (e.g., leptin, insulin and ghrelin) have carriers or transporters that facilitate their access to deeper brain sites (23, 25–27).

B. Metabolic Cues With Primary Targets in the CNS

The metabolic control of the reproductive system is ultimately exerted by the modulation of the activity of GnRH neurons and changes in the pulsatile release of GnRH into the portal vessels of the median eminence. Altered GnRH pulsatile release (exacerbated or insufficient) disrupts gonadotropin secretion and, consequently, sex steroid production impairing gametogenesis and sexual behavior (294, 487). Metabolic cues modulate GnRH neuronal activity, GnRH release, and sexual behavior via distinct mechanisms and brain circuitry.

1. Adipocyte-derived hormones: leptin

The adipose tissue secretes hormones (adipokines) that are key for whole body physiology (4, 167, 431). The imbalance of these hormones due to excess, absent, or dysfunctional adipocytes disrupts homeostasis and has deleterious consequences for individual health. Several adipokines directly affect the reproductive function. However, only leptin has been implicated in reproductive control via actions in the CNS.

Seminal findings showed that loss-of-function mutations in leptin (Lep/LEP) or leptin receptor (Lepr/LEPR) genes cause metabolic disorders and infertility in rodents and humans (88, 92, 147, 218, 222, 513). These autosomal recessive mutations were initially identified in mice at the Jackson Laboratories (named ob/ob or Lep^ob for obese and db/db or Lepr^db for diabetes) (218, 222). Mice homozygous for Lep or Lepr mutations are morbidly obese due to hyperphagic behavior, decreased energy expenditure, and multiple neuroendocrine dysfunctions (92, 218, 222). At peripubertal stages, the reproductive organs of mutant mice resemble those of wild types at the same developmental stage, but sexual maturation is arrested (28, 75, 317). Leptin-deficient mice have a typical distribution of GnRH somata and terminals and high hypothalamic GnRH content but low circulating levels of gonadotropins (30, 437). Exogenous GnRH administration induces LH secretion, suggesting that leptin-deficient mice have functional gonadotropes but deficient GnRH release (30, 488). Indeed, in vitro and in vivo studies have shown that leptin acts primarily in the brain by stimulating GnRH secretion (140, 493, 509).

The administration of leptin at doses that do not decrease body weight advances the onset of puberty in female wild-type mice (5, 168). In prepubertal female rats, leptin increases the secretion of gonadotropins and the expression of ovarian steroidogenesis enzymes, contributing to pubertal development (12). In doses associated with suppressed food intake and decreased body weight, leptin does not advance the age of puberty onset. However, control pair-fed mice showed delayed sexual maturation, suggesting that higher leptin levels in leptin-treated animals neutralized the delay in puberty caused by a state of negative energy balance (76, 77, 187). Moreover, the overexpression of leptin in transgenic skinny mice induced early puberty onset with uterine and ovarian development, indicative of accelerated activity of the HPG axis (510). The skinny female mice also showed early reproductive senescence, indicating that hyperleptinemia induces premature reproductive failure. Whether this mechanism explains the advanced puberty in obese girls or the dysfunctional activity of the HPG axis in obese individuals remains to be directly addressed (10, 43, 59, 133, 138, 227, 442).

Circulating levels of leptin correlate with stored white adipose tissue mass and follow a circadian pattern in rodents and primates (7, 94, 275, 276, 284, 513). However, during acute fasting or calorie restriction, unknown mechanisms cause leptin levels to fall before the depletion of adipocyte stores. First-order neurons perceive the decrease in leptin levels as a state of energy insufficiency. As a result, counter-regulatory responses are triggered to retain the energy necessary for basic physiological needs (8, 73, 151). Examples of adaptive responses are decreases in thermogenesis and locomotor activity, inhibition of the thyroid axis, and activation of the hypothalamo-pituitary-adrenal (HPA) axis (6, 8, 9, 72, 197, 268, 275, 323). Because of the high energetic costs of reproduction, the HPG axis is inhibited during fasting or in conditions of low leptin levels. Rodents and primates in a negative energy balance exhibit decreases in pulsatile LH secretion, sex steroids, and fertility (60, 62, 63, 140, 292). In several species, following distinct experimental paradigms, leptin administration during states of negative energy balance increases LH pulsatile release, restores female cyclicity, and improves fertility (8, 69, 181, 324, 478). The effect of leptin to increase LH was not observed in calorie-restricted rats under constant leptin administration using osmotic pumps (453). Because of the pulsatile nature of neuroendocrine control and the circadian patter of leptin secretion, it is not clear if the lack of fluctuations in circulating levels disrupted the physiological effects of leptin on the reproductive axis. In women with hypothalamic amenorrhea due to extreme exercise and weight loss, leptin increases LH pulsatility, sex steroid levels, ovarian volume, and the number of maturing follicles (73, 276, 309, 476, 481). Therefore, leptin serves as a signal of energy sufficiency for sexual maturation and maintenance of fertility in humans and rodents. In seasonal breeders, however, the role of leptin becomes secondary, as the main environmental signals are conveyed by changes in photoperiod (399, 400).

Leptin’s actions on reproductive function are mediated by LepR expressed in the CNS (11, 91, 118). Studies using different approaches have determined that these effects are exerted by neurons upstream of GnRH cells (127, 280, 371). Since LepR expression only in calmodulin-dependent protein kinase IIα (CaMKIIα) neurons can maintain reproductive function, the forebrain is believed to be the main player in leptin’s effect on the HPG axis (371). Work from several groups has identified the PMV, the medial preoptic area (MPO), and the ARH as the main targets of leptin action in the neuroendocrine reproductive axis (33, 125, 127, 137, 272, 280, 290, 345, 412) (FIGURE 4).

FIGURE 4.

Schematic illustration showing coronal sections of hypothalamic sites directly targeted by leptin in mice. The neuronal populations associated with leptin action in metabolic and reproductive control, i.e., the medial preoptic area (MPO) and the arcuate nucleus (ARH), are represented in blue. Green highlights the sites associated with distinct aspects of the metabolic control, including the dorsomedial subdivision of the ventromedial nucleus (VMH), the dorsomedial nucleus (DMH), and the lateral hypothalamic area (LHA). In orange is represented the ventral premammillary nucleus (PMV), associated with reproductive control. 3V, third ventricle; ac, anterior commissure; f, fornix; LV, lateral ventricle; MS, medial septum; opt, optic tract.

The role of the PMV as a primary site of leptin action in reproductive control was demonstrated by studies using lesions and viro-genetic manipulation in rodents (125, 127). Obese and infertile LepR-null female mice showed pubertal development and improved fertility after the reexpression of LepR in PMV neurons. Thus leptin action in the PMV is a sufficient stimulus for the reproductive axis (127). Excitotoxic lesions of the PMV of rats decreased the activation of GnRH and AVPV neurons and the circulating levels of LH and estradiol during proestrus. They also blunted the ability of leptin to increase LH during fasting (125). In ob/ob mice, PMV lesions caused a significant delay in pubertal maturation following leptin administration. Because puberty was delayed but not arrested, these findings indicate that leptin action in the PMV is not required for sexual maturation and that alternative pathways exist.

PMV LepR neurons directly innervate GnRH and kisspeptin neuronal populations. Approximately 75% of these neurons depolarize after acute leptin administration (49, 127, 272, 280, 386, 490). The classic excitatory amino acid glutamate (51, 288) and the gaseous neurotransmitter nitric oxide (NO) are the main neurotransmitters in PMV projections (127, 129, 272) (See note added in proof).

The effects of NO on GnRH secretion are well established. Neuronal NO synthase (nNOS; Nos1 gene) knockout mice are infertile. In addition, leptin-induced GnRH/LH secretion is blocked following the global deletion or inhibition of NO production (33, 87, 188, 235, 376, 477, 508, 509). Leptin administration increases the activity of nNOS, subsequently increasing NO production in LepR neurons of the PMV and the MPO (33, 129, 347). In the MPO, the inhibition of NO prevents leptin from inducing LH secretion in fasted mice (33). Notably, the deletion of LepR from nNOS cells disrupts metabolic function and pubertal development (271). Together, these findings indicate that the PMV and the MPO are two primary sites where leptin may promote reproduction via NO neurotransmission (FIGURE 5).

FIGURE 5.

Schematic illustration showing coronal sections of brain sites and neural pathways associated with leptin action in reproductive control in mice. Neurons in the medial preoptic area (MPO) and in the ventral premammillary nucleus (PMV) coexpress nitric oxide synthase (nNOS) and glutamate and alter gonadotropin releasing hormone (GnRH) neuronal activity via direct projections or innervation of kisspeptin neurons. Neurons in the arcuate nucleus (ARH) expressing neuropeptide Y (NPY), agouti-related protein (AgRP), and GABA as well as those expressing proopiomelanocortin (POMC) project to kisspeptin. Thick purple lines in the median eminence (ME) represent GnRH terminals at the portal circulation. ac, anterior commissure; AVPV, anteroventral periventricular nucleus; f, fornix; LV, lateral ventricle; ME, median eminence; MS, medial septum; OVLT, vascular organ of the lamina terminalis.

The ability of leptin to facilitate lordosis is reduced when NO production is blocked (61, 172, 180, 291, 470). The brain pathways involved in this effect are not known, but strong interconnections between either the PMV or the MPO and the VMHvl, a key site in the control of lordosis in female rodents, have been reported (66, 67, 386). The physiological relevance of these reciprocal connections requires further investigation. Because leptin’s behavioral effects require GnRH signaling, NO and leptin may jointly alter sex steroid levels via actions on GnRH neurons, indirectly affecting behavioral responses.

The melanocortin system in the ARH is an alternative site through which leptin modulates reproduction. A distinct subset of first-order neurons expressing agouti-related protein (AgRP), neuropeptide Y (NPY), and GABA, as well as another expressing proopiomelanocortin (POMC) and predominantly glutamate have well-described actions in metabolism and leptin physiology (139, 189, 254, 489, 498, 502). The posttranslational cleavage of POMC generates melanocortins [e.g., α-melanocyte stimulating hormone (α-MSH)], adrenocorticotropic hormone (ACTH), opioids (β-endorphin), and lipotrophins (332). Of those, the neuropeptides α-MSH and β-endorphin are the main players in reproductive control. While previous studies have shown that the deletion of LepR from neurons expressing AgRP, POMC, or both does not disrupt reproduction (24, 462), the ablation of AgRP/NPY neurons or the blockade of melanocortin receptors improved fertility in leptin-signaling deficient mice (141, 224, 412, 502). By suppressing the activity of AgRP neurons, leptin can induce puberty and improve fertility in mice. However, leptin action on AgRP neurons is sufficient but not required for reproductive function, as deletion of LepR from AgRP neurons does not preclude reproduction (137, 144). The effects of AgRP/NPY neurons seem to be attained mainly by restraining kisspeptin neuronal activity via GABAergic neurotransmission (345). This is in line with studies showing that the deletion of LepR from GABAergic neurons delays sexual maturation and disrupts the estrous cycle in mice (297, 515) (FIGURE 5). On the other hand, mice lacking both LepR and insulin receptor (IR) in POMC cells (LepR/IR^POMC mice) exhibited systematic insulin resistance as well as reduced fertility, suggesting cooperative actions by leptin and other metabolic signals.

The melanocortin system may also act downstream of leptin to affect sexual behavior. Male LepR/IR^POMC mice showed reduced α-MSH production and decreased reproductive success due to dramatic changes in sexual behavior (207). In pubertal rats, melanocortin receptor agonists activate the HPG axis by stimulating kisspeptin release (290). Kisspeptin neurons of mice coexpress melanocortin receptors, and the inhibition of KNDy neurons using designed receptors exclusively activated by designed drugs (DREADDs technology) blocks the ability of melanocortins to stimulate the reproductive axis (290). The disruption of the melanocortin system in humans, including POMC and MC4R genes, has been associated with obesity with virtually no effect on reproductive function (146, 459). On the other hand, individuals with the deletion of the prohormone convertase 1 (PCSK1) gene, an enzyme that processes distinct prohormones including POMC, show obesity and hypogonadotropic hypogonadism (432, 463). Whether this phenotype is associated with the posttranslational processing of POMC or other prohormones is undefined.

The melanocortin system has a relevant role in sexual arousal and penile erection in rodents and humans (68, 457, 464, 467). High rates of sexual dysfunction including disorders of libido, decreased sexual desire, and reduced intercourse satisfaction have been reported in obese and type 2 diabetes (T2D) patients (344, 389, 412), in whom hypothalamic leptin and insulin resistance are likely. Thus leptin signaling in POMC neurons may play a role in supporting the necessary release of α-MSH to maintain reproductive behavior.

2. Pancreatic islet-derived hormones: insulin

The pancreatic islets produce hormones with key actions in homeostasis, including insulin, glucagon, and somatostatin. While the effects of circulating glucagon and somatostatin in the central control of the reproductive function are understudied, insulin is well recognized as having a substantial impact on reproduction.

Insulin is a polypeptide hormone that islet beta cells secrete in response to high blood glucose levels in the short term and elevated body adiposity in the long term (466, 496). Insulin regulates the metabolism of carbohydrates and fat by promoting glucose uptake from the circulation to skeletal muscle, liver, and adipose tissues. IR are found at all levels of the HPG axis, allowing insulin to modulate reproductive function and behavior in a tissue-specific manner. In the ovaries of animal models and women, insulin potentiates LH action by increasing steroidogenesis (3, 327, 503).

Although apparently not required for glucose uptake by neurons, IR are located throughout the brain, with the highest concentrations in the olfactory bulb, cerebral cortex, cerebellum, hippocampus, and hypothalamus (195, 293, 364, 484). In rodents, the highest concentrations of IR are found in the MPO, ARH, and VMH (56, 293).

Circulating insulin levels correlate positively with body weight and adiposity, indicating that insulin resistance increases with weight gain (497). Intracerebroventricular or intranasal application of insulin decreases body weight and food intake in rats, mice, and primates, including humans (55, 74, 190, 496). Additionally, the blockade of insulin signaling in the hypothalamus increases food intake and body weight (336).

Insulin also plays a significant role in the neuroendocrine control of the reproductive function (56, 186, 336). Brain-specific deletion of insulin receptors results in diet-sensitive obesity and HPG axis dysregulation (56). In obese female mice, insulin increases GnRH activity and LH secretion, potentially contributing to obesity-induced reproductive dysfunction (123). However, the specific cellular targets associated with insulin action on reproductive neuroendocrine axis have not been determined. Studies using conditional deletions of IR in distinct neuronal populations (CaMIIKα, AgRP, POMC, GnRH, and kisspeptin) have not been very informative, as these mouse models failed to display changes in reproductive phenotype or the GnRH network dysregulation (124, 143, 145, 207, 252). However, redundant pathways may exist, and the deletion of IR from one population of neurons early during development may result in compensation from other cell populations or metabolic pathways. Alternatively, the hypothalamic hypogonadism observed in brain-specific IR knockout mice may have resulted from the absence of insulin signaling in nonneuronal cells (e.g., glia). This hypothesis is in line with findings showing that obesity and T2D induce signs of hypothalamic inflammation and mitochondrial dysfunction associated with cellular insulin resistance (444, 460). Notably, IR in astrocytes assist glucose transport via the BBB, thus altering the signals of glucose availability to relevant neurons (166). Whether this function also affects reproduction is unknown.

3. Gastrointestinal tract and accessory organs: ghrelin, GLP1, FGF21, and IGF1

Enteroendocrine cells in the gastrointestinal (GI) tract and hepatocytes communicate signals of immediate energy availability to the reproductive system. These cells function as nutrient sensors that respond to GI luminal content (the amount and composition of food) and release hormones to inform downstream targets about the immediate nutritional state. These nutritional signals reach the CNS via direct release into the circulation (hormones) or via afferent neurons to the nucleus of the solitary tract (39). Some of these hormones modulate reproductive physiology.

a) ghrelin.

Ghrelin is secreted by gastric mucosal cells during conditions of undernutrition (206, 318). Originally recognized for its growth hormone-stimulating action, ghrelin has potent effects on energy balance independent of the somatotropic axis. Ghrelin promotes the storage of lipids as fat, induces glucagon release, and suppresses insulin secretion and sensitivity (319). In addition to its effects on the growth axis and energy balance, ghrelin also modulates reproductive function (319, 439). In mice, the functional ghrelin receptor GHSR1a is widespread in the hypothalamus, with high expression in the AVPV, MPO, and mediobasal hypothalamus (156, 514). No reproductive phenotype has been described in mice with the global deletion of ghrelin (Ghrl gene knockout) (433). However, compensatory mechanisms may be in place, as the inhibitory actions of ghrelin on reproduction have been shown in several species including sheep, monkeys, and humans (439). For example, the intracerebroventricular or peripheral administration of ghrelin suppresses gonadotropin secretion in rats. Likewise, ghrelin inhibits GnRH release from hypothalamic explants (149, 150, 161).

The neural basis for ghrelin action in the HPG axis is not completely known, but studies have suggested that the kisspeptin neurons are a potential target. GHSR1a is expressed in AVPV/PVpo kisspeptin and KNDy neurons in mice (156). Ghsr expression is regulated by estradiol, i.e., higher expression and colocalization was observed in KNDy neurons of ovariectomized mice treated with estradiol. The higher colocalization was not caused by changes in Kiss1 mRNA expression, as the procedure was performed in Kiss1-Cre mice crossed with GFP reporter under the R26 promoter, and therefore GFP immunoreactivity was stable across estrous cycles and under different sex steroids milieu. Contrary to expectations, KNDy neurons depolarize in response to ghrelin (156). Because higher Kiss1/GHSR1a coexpression is observed when Kiss1 mRNA levels are low (in circumstances of high circulating E2) (425), it is possible that the KNDy neurons preferentially release neuropeptides or neurotransmitters other than kisspeptin when activated by ghrelin. While this model could plausibly explain the unexpected findings, it requires verification.

Ghrelin may also inhibit reproduction by increasing the activity of AgRP/NPY neurons (100). As discussed in previous sections, the stimulation of this pathway may decrease the activaton of the HPG axis by inhibiting kisspeptin neurons (345) (FIGURE 6). In female rats, peripheral ghrelin had no effect on ARH Kiss1 mRNA levels, but it decreased Kiss1 expression in the AVPV (153). Because Ghsr is not expressed in the AVPV of rats (514), upstream interneurons (e.g., AgRP neurons) may have caused this effect (345).

FIGURE 6.

Schematic illustration showing coronal sections of brain sites and neural pathways targeted by hormones from the gastrointestinal tract and liver with action in reproductive control in mice. Ghrelin receptor is coexpressed in agouti-related protein (AgRP) and kisspeptin neurons. Glucagon-like peptide 1 (GLP1) modulates Kiss1 expression in the arcuate nucleus (ARH). Fibroblast growth factor 21 (FGF21) acts on FGF receptor 1/βKlotho expressed in vasopressin neurons of the suprachiasmactic nucleus (SCN) upstream of kisspeptin neurons of the anteroventral periventricular nucleus (AVPV). Insulin-like growth factor 1 (IGF1) targets gonadotropin releasing hormone (GnRH) neurons and kisspeptin neurons in the AVPV. Thick purple lines in the median eminence (ME) represent GnRH terminals at the portal circulation. ac, anterior commissure; f, fornix; LV, lateral ventricle; ME, median eminence; MS, medial septum; opt, optic tract; OVLT, vascular organ of the lamina terminalis; ox, optic chiasm; PVH, paraventricular nucleus of the hypothalamus.

In male rats, ghrelin increases the latency to first mount, intromission, ejaculation and the length of postejaculatory intervals. Coadministration of ghrelin with the ghrelin receptor antagonist (d-Lys3)-GHRP6 attenuated ghrelin’s effects (19). Ghrelin may also inhibit female sexual behavior and receptivity during fasting conditions, when circulating ghrelin levels are high. Ovariectomized steroid-primed females subjected to acute food restriction and ghrelin administration showed decreased sexual receptivity. Administration of a ghrelin receptor antagonist to these females restored sexual behavior (40). However, chronic ghrelin administration did not reduce sexual receptivity. The mechanisms underlying the differential effects of acute and chronic ghrelin treatments are not completely known, but a compensatory response involving receptor downregulation is a potential explanation (40). Whether ghrelin exerts its effects on sexual behavior via direct actions on brain nuclei associated with behavioral control is unknown.

Notably, one feature common to individuals with Prader-Willi syndrome (PWS) is high levels of circulating ghrelin. PWS affected individuals are obese, hyperphagic, and show lack or delay in pubertal development and low gonadotropin levels (415). It is possible that high ghrelin levels in PWS have an effect to inhibit the HPG axis contributing to the disrupted fertility observed in affected individuals. Further studies are necessary to test this model.

b) glp1.

GLP1 is an incretin member of the structurally related glucagon family of peptide hormones, with actions in diverse physiological systems including feeding, gastric motility, and cardiovascular function (130). GLP1 is primarily synthesized from the proglucagon gene by the ileum L cells in response to food intake (29, 130). Studies have shown that the peripheral or central administration of a GLP1 receptor agonist acts on cognate receptors in the CNS to suppress feeding and decrease body weight (142).

GLP1 also modulates the reproductive physiology of rodents (201, 342). The acute administration of GLP1 increases the amplitude of the preovulatory LH surge, and chronic GLP1 treatment of prepubertal rats advances the onset of puberty (342). GLP1 also enhances LH secretion in male rats and promotes GnRH release in hypothalamic explants (31). Similarly, mice with a global deletion of GLP1 receptors show delayed onset of puberty and low gonadal weight (283). The sites of GLP1 action in reproduction are unclear, but part of these effects may involve kisspeptin neurons (200). In mice, GLP1 immunoreactive fibers are in close apposition with KNDy neurons, and electrophysiological recordings showed that GLP1 receptor agonists depolarize and increase the firing rate of those neurons (200) (FIGURE 6). GLP1 is also expressed in neurons of the brain stem (263, 468), but whether its effects on reproductive physiology are mediated by neuronal or humoral actions have not been demonstrated.

c) fgf21.

FGF21 is secreted from the liver, pancreas, and white adipocytes and acts as a hormone in several endocrine systems, with broad actions in metabolic control (247). FGF21 signals through the FGF receptor 1 (FGFR1) complexed with a membrane protein or co-receptor named βKlotho. Whereas FGFR1 is widespread, βKlotho has a restricted distribution. Other FGFs also bind to FGFR1; therefore, the sites of βKlotho expression are likely key for FGF21 actions. After a prolonged fast, FGF21 levels rise, inducing hepatic gluconeogenesis and fatty acid oxidation (22). Because of the effects of negative energy balance on fertility, studies have assessed the direct role of FGF21 as a signal of energy insufficiency to the HPG axis. Mice exposed to FGF21 show infertility secondary to a negative energy balance, an effect attributed to actions in the hypothalamic suprachiasmatic nucleus (SCN) upstream of the vasopressin-kisspeptin pathway (343). βKlotho is highly expressed in SCN vasopressinergic neurons, suggesting that FGF21 binding may inhibit the LH surge by disrupting circadian control (343) (FIGURE 6). Mice lacking βKlotho in the SCN are protected against the inhibitory effect of FGF21 on reproductive function. However, following a ketogenic diet, mice displayed high levels of circulating FGF21 but remained fertile, and the intracerebroventricular administration of FGF21 had no effect on the reproductive axis (419). Loss-of-function mutations in the KLB gene (coding for βKlotho) in humans cause congenital hypogonadotropic hypogonadism, and mice lacking the Klb gene show delayed puberty, deficient estrous cycles, and subfertility (504). These effects seem to be caused by a lack of response of GnRH neurons to FGF21 stimulation. Together, these findings suggest that FGF21 affects GnRH physiology, but whether it conveys signals of energy sufficiency remains to be independently assessed.

d) igf1.

IGF1, also called somatomedin C, is a protein hormone that in humans is encoded by the IGF1 gene (213, 225). IGF1, IGF2, and insulin are members of the same family of insulin-like peptides (97). IGF1 plays an important role in childhood growth and continues to have anabolic effects in adults (237). A large proportion of IGF1 is synthesized in the liver and released into the plasma, but IGF1 is also expressed in many tissues at low levels. Studies with rat models of low circulating IGF1 have shown that brain IGF1 is regulated independently of plasma IGF1 levels (2).

IGF1’s primary action is mediated by binding to its specific tyrosine kinase receptor, the IGF1 receptor (IGF1R), which is expressed in a wide range of tissues, including the brain. Both IR and IGF1R exist as heterotetramers composed of two dimers (alpha and beta) (325), although hybrids exist consisting of combinations of IR and IGF1R subunits (36). Previous evidence showed that insulin and IGF1 bind to their noncognate receptor with an order of magnitude lower than their cognate receptors (173, 246) and to hybrid receptors with varying affinities (36, 393). The overlap between the signaling pathways for insulin and IGF1 indicate that the temporal and spatial expression of the receptors underlies the specificity of the responses.

IGF1 of both peripheral and central origin can interact with hypothalamic IGF1R. Levels of IGF1 in the brain have been shown to play a role in puberty, estrous cyclicity, the preovulatory gonadotropin surge, and the decline of reproductive neuroendocrine function during reproductive senescence (109, 110, 124, 308). IGR1 and IGF1R are both found in GnRH neurons (108). Precocious puberty was observed in female mice following peripheral treatment with IGF1 (113, 124), although no change was seen in males (124). Intracerebroventricular infusion of IGF1 stimulates secretion of GnRH in prepubertal female rats and advances the onset of puberty (209, 210). Nevertheless, GnRH expression was not altered in IGF1-treated rats as assessed by RNase protection assay analysis of RNA from hypothalamic tissue (308). Conversely, brain IGF1R antagonism impairs estradiol positive feedback, reduces estradiol- and progesterone-dependent LH surges, and disrupts estrous cycling independent of changes in body weight or food intake (434, 447). Under estradiol positive feedback conditions, intracerebroventricular infusion of the IGF1R antagonist JB-1 in young adult females reduces GnRH neuron activation, as determined by the percentage of GnRH/c-Fos coexpressing cells (330). These findings suggest that IGF1R signaling functions as an energy indicator required for normal hypothalamic control of reproduction. Mice lacking the IGF1R in GnRH neurons (IGF1R^GnRH) exhibited a delay in the onset of puberty, but showed normal reproductive function as adults (124). These results are similar to those observed when IR was specifically deleted in the mouse kisspeptin neurons (IR^∆Kiss) (370) underscoring the diversity of metabolic signals that likely contribute to the pubertal process.

IGF1R signaling in neurons is also important for neuronal progenitor proliferation, neurite outgrowth, and synaptogenesis necessary for interneuronal communication (334). Reduced levels of hypothalamic IGF1 have been proposed to underlie the decline in neuroendocrine function associated with reproductive aging. During pubertal progression, hypothalamic IGF1 levels are increased, while a significant reduction in hypothalamic IGF1 was reported in older rats undergoing reproductive senescence (308). Moreover, IGF1 signaling was required for proper timing and magnitude of the estradiol induction of the GnRH surge (434, 448). Indirect effects on GnRH release can also be mediated by kisspeptin neurons (FIGURE 6). Both centrally perfused IGF1 and peripherally injected IGF1 stimulated the kisspeptin neurons of the AVPV (211). These data may also provide an anatomical explanation for the sexual dimorphism in the response of mice to IGF1 since males have limited AVPV kisspeptin expression.

4. Oxidizable fuels and metabolites

The brain also monitors the availability of energy using metabolic substrates and the oxidation of fuels. Among them, glucose, free fatty acids, and amino acids act as immediate signals of the nutritional state to the CNS (399, 407). Sensors for these metabolic substrates and oxidizable fuels reside in areas also known to control reproduction.

a) glucose.

Glucose is a primary fuel source for the CNS and is fundamental for neuronal function (13). Neuronal glucose-sensing mechanisms allow the brain to constantly monitor intracellular glucose levels and, in turn, control energy homeostasis in peripheral organs (273). Glucose also acts as a signaling molecule in glucose-sensitive neurons that may be subdivided into glucose-excited (GE) and glucose-inhibited (GI) neurons (233, 474). The firing rate of GE neurons increases and that of GI neurons decreases as glucose levels rise (134). Most GE neurons express anorexigenic peptides (e.g., α-MSH from POMC neurons), whereas GI neurons express appetite-stimulating peptides (e.g., NPY and AgRP) (220, 322).

Changing glucose levels are a key signal to the neuroendocrine reproductive axis. For example, treatment with a competitive inhibitor of intracellular glucose utilization (i.e., 2-deoxy-d-glucose or 2DG, and 5-thio-glucose or 5TG) inhibits pulsatile LH secretion and pulse frequency (245, 259, 320, 337). Administration of glucose to hypoglycemic rats restores LH secretion (382). The effects of glucose on the modulation of the HPG axis may engage three potential pathways: 1) actions in neurons of the mediobasal hypothalamus (including the melanocortin system) (220, 322), 2) glucose sensing in the area postrema and in catecholaminergic neurons of the caudal medulla (71, 219, 320), and 3) direct actions in GnRH neurons (383, 512) (FIGURE 7).

FIGURE 7.

Schematic illustration showing coronal sections of brain sites and neural pathways responsive to metabolites and oxidizable fuels with action in reproductive control in mice. Glucose directly affects the proopiomelanocortin (POMC) and agouti-related protein (AgRP) neurons of the arcuate nucleus (ARH), the area postrema (AP), and the catecholaminergic neurons of the caudal medulla. These sites innervate the paraventricular nucleus of the hypothalamus (PVH), the main output to behavioral control (via oxytocin and MC4R neurons), and neuroendocrine output (via corticotropin releasing hormone/CRH and oxytocin neurons). Glucose and free fatty acid (FFA) potentially alter gonadotropin releasing hormone (GnRH) neurons activity. AMP kinase (AMPK) and mTOR change kisspeptin response to metabolic challenges. Thick purple lines in the median eminence (ME) represent GnRH terminals at the portal circulation. A1/C1 and A2, noradrenergic and adrenergic neuronal groups in the caudal medulla; ac, anterior commissure; ARH, arcuate nucleus; AVPV, anteroventral periventricular nucleus; E, epinephrine; f, fornix; LV, lateral ventricle; ME, median eminence; MS, medial septum; NE, norepinephrine; NTS, nucleus of the solitary tract; ox, optic chiasm; PP, posterior pituitary gland; PVH, paraventricular nucleus of the hypothalamus.

Ablation of hypothalamic projecting catecholaminergic neurons of the caudal medulla (A1/C1 and A2 groups) disrupts the adaptive responses of the reproductive axis to low glucose levels (219). This finding was obtained using selective deletion of catecholaminergic neurons that innervate the paraventricular nucleus of the hypothalamus (PVH), suggesting that the medulla-PVH projection plays an important role conveying signals of glucose levels to the reproductive neuroendocrine axis. However, as discussed by the investigators, deletion of A1/C1 and A2 groups of neurons also decreased cathecolaminergic innervation of ARH and MPO. Therefore, the lack of response to changes in glucose levels may be associated with decreased adrenergic/noradrenergic innervation of any one or of all the hypothalamic sites compromised by the procedure.

The PVH houses the corticotropin releasing hormone (CRH) neurons of the HPA axis. In general, stressors of different origins, including those triggered by a negative energy balance, activate the HPA axis by inducing the secretion of CRH into the hypophysial portal system (8, 111, 285, 456). The neural path or mechanisms by which the HPA axis is activated in conditions of nutritional stress are not completely known, but changes in glucose and leptin levels seem to exert a crucial role. Paradigms of 2DG administration to decrease glucose utilization activate the HPA axis and inhibit LH secretion (456, 479). Pharmacological inhibition of CRH signaling in 2DG-treated female rats normalizes LH levels, potentially via direct actions on GnRH neuronal activity (355, 456). Leptin administration to fasted mice blocks the activation of the HPA axis and the suppression of the HPG axis (8, 197).

A subpopulation of GnRH neurons has its firing rate reduced following exposure to low concentrations of glucose (383, 512). This effect seems to be mediated in part by ATP-sensitive potassium channels (K_ATP) and the enzyme glucokinase. When transported into cells, glucose serves as a substrate for ATP production via the actions of glucokinase. High levels of intracellular ATP block K_ATP channels, decreasing potassium efflux and causing depolarization of GE neurons. Alternatively, changing levels of glucose may also alter the intracellular AMP-to-ATP ratio, modulating the enzyme AMP kinase (AMPK). Low extracellular glucose reduces intracellular ATP production, raising the AMP-to-ATP ratio and AMPK activation (194). AMPK acts at the intracellular level to inhibit ATP utilization and stimulate ATP production (430). In female rats, the intracerebroventricular infusion of an AMPK activator disrupted estrous cycles, and the acute application of an AMPK agonist into hypothalamic slices hyperpolarizes GnRH neurons (101). Likewise, AMPK antagonists attenuate glucose sensing in GnRH neurons (383, 384).

The non-neuronal actions of AMPK may also play a role in GnRH suppression. Malnutrition-induced AMPK activation of glial cells suppresses GnRH/LH release (310). Interestingly, AMPKα1-knockout mice are subfertile, but exhibit no apparent metabolic phenotype, indicating that subfertility in these mice is not secondary to metabolic dysregulation (230). However, the conditional deletion of AMPKα2 in Kiss1 neurons causes neither metabolic nor reproductive deficits but does disrupt the ability of the female mice to adapt to metabolic challenges (451).

b) free fatty acids.

Although glucose is the primary energy source for the CNS, the oxidation of free fatty acids (FFA) may also provide a fraction of energy to the brain (135, 346). The limited use of FFA compared with glucose is likely due to the low β-oxidation by brain mitochondria (403). Nevertheless, the CNS may perceive changes in FFA levels and alter physiology (152, 261). In Syrian hamsters, availability of FFA (via fat oxidation) maintains the reproductive cycle in food-restricted females (401, 402). While minimal data have connected these sensors to the CNS in reproductive regulation, one study found that FFA change GnRH mRNA expression in a GnRH-synthesizing cell line (452) (FIGURE 7). Additional studies will be necessary to assess the physiological relevance of these findings.

c) amino acids.

Changes in dietary protein influence feeding behavior. Rats placed on a low-protein diet exhibit increased food intake and AgRP expression. Intracerebroventricular injection of an amino acid mixture, or just leucine, suppresses food intake (315). Amino acids (l-leucine) along with adequate levels of ATP activate the mammalian target of rapamycin (mTOR) pathway, which induces protein synthesis and growth (193, 215, 398). A loss-of-function mutation in the Mtor gene is embryonically lethal, whereas increased activity of mTOR induces metabolic dysfunctions including obesity and diabetes (164, 199, 215). The mTOR pathway also controls the HPG axis and the onset of puberty (379, 381). The blockade of mTOR signaling by rapamycin inhibits puberty, decreases LH and estradiol levels, and causes ovarian and uterine atrophy. The central activation of mTOR stimulates LH secretion. These actions may result from the mTOR control of KNDy neurons, since the inhibition of mTOR preferentially decreases Kiss1 transcription levels in the ARH (379) (FIGURE 7).

C. Integrative Sites and Neuronal Networks

The metabolic control of reproductive function is exerted ultimately by changing the activity of GnRH neurons. This effect requires direct actions on GnRH somata and terminals or the modulation of kisspeptin neurons (FIGURE 8). Most of the first-order neurons that sense changes in energy availability are located in the ARH. This is not surprising as the ARH neighbors a circumventricular organ highly relevant for neuroendocrine control, the median eminence. Therefore, most of the ARH neurons are located outside the BBB.

FIGURE 8.

Summary of the neuronal network associated with the metabolic control of reproductive function described in rodents. Metabolic cues are listed in the left column, specific sites in the central nervous system (CNS) are distributed in the middle, and the endocrine and behavioral outputs are presented in the right. Black lines represent well-defined networks, and dashed lines represent proposed, still undefined, projections. Key sex steroid targets are represented in green arrows. AgRP, agouti-related protein; AP, area postrema; ARH, arcuate nucleus; AVP, vasopressin; AVPV, anteroventral periventricular nucleus; CRH, corticotropin releasing hormone; E, epinephrine; f, fornix; GnRH, gonadotropin releasing hormone neurons; KNDy, kisspeptin neurokinin B dynorphin; LV, lateral ventricle; MC4R, melanocortin 4 receptor; MPO, medial preoptic area; NE, norepinephrine; nNOS, nitric oxide synthase; NPY, neuropeptide Y; POMC, proopiomelanocortin; PVH, paraventricular nucleus of the hypothalamus; SCN, suprachiasmatic nucleus; PMV, ventral premammillary nucleus; VMHvl, ventrolateral subdivision of the ventromedial nucleus of the hypothalamus.

ARH neurons coexpress the orexigenic neuropeptides NPY and AgRP, both critical mediators of energy homeostasis (189, 489, 498). NPY/AgRP neurons are stimulated by signals of low energy and are inhibited in conditions of excess energy (100, 252). The suppression of basal LH levels by fasting fails in NPY knockout mice (208). However, pharmacological studies in rodents and primates (106, 494) have suggested that the NPY has a dual action on the HPG axis according to sex steroids milieu. NPY has suppressive effects on LH pulse amplitude and frequency when sex steroids are low and a robust activational effect following steroid replacement in gonadectomized rats (38, 106, 301). Moreover, LH surges are attenuated after NPY immunoneutralization or Y1 receptor blockade, and NPY-knockout female mice show reduced LH surges (505). AgRP/NPY neurons modulate reproduction by acting on both GnRH and kisspeptin neurons (224, 380, 412, 501, 502). The effects of NPY on GnRH neurons rely on the specific receptors recruited by the ligand. NPY inhibits ~10% of GnRH neurons through Y1 receptors but stimulates 25% of GnRH neurons through Y4 receptors (380). Its higher affinity for the Y1 receptor suggests an inhibitory net effect upon GnRH neurons. However, the genetic deletion of the Y4 receptor in obese infertile ob/ob mice restores fertility in males and improves the reproductive phenotype of females (392). Whether NPY modulates the HPG axis via Y4 receptors apart from those receptors on GnRH neurons has not been demonstrated. NPY/AgRP neurons also innervate kisspeptin cells. This input seems to be exerted primarily via melanocortin receptors or by GABAergic neurotransmission (20, 345). According to this model, NPY neurons are activated under conditions of negative energy balance, leading to an increase in NPY release in the ARH and AVPV.

ARH POMC neurons are also highly responsive to metabolic cues, including leptin, insulin, and glucose (207). In conditions of negative energy balance, such as fasting, POMC mRNA expression drops, and the opposite is seen when energy is abundant (445, 489, 498). POMC neurons in the ARH coexpress the neuropeptide cocaine- and amphetamine-regulated transcript (CART) and project to brain sites associated with metabolic and reproductive control (139, 386, 454).

The melanocortin system modulates reproductive function via melanocortin receptors (MC3R and MC4R). The MC4R-knockout mice are morbidly obese, but their reproductive deficits are mild (397). However, females show reduced ovulation and develop subfertility later in life (21, 290, 454). Whether this effect is caused by increased adiposity has not been demonstrated. A direct stimulation of GnRH neurons by α-MSH via MC3R and MC4R signaling has been demonstrated (380), suggesting the GnRH neurons are the melanocortin preferential targets. However, melanocortin receptors are also expressed in kisspeptin neurons (21, 104), and POMC/CART neurons directly regulate kisspeptin output (21, 290, 454). Due to the sensitivity of POMC/CART neurons to metabolic cues, the POMC-Kiss1-GnRH circuit may function as a crucial neuronal path by which signals of energy sufficiency modulate the HPG axis (290).

Obese MC4R null males show decreased motivation to engage in sexual activity as well as reduced erectile function and take longer to reach ejaculation (148, 409, 464). Blockade of the MC4R reduces mounting and increases latency to intromission, and the administration of α-MSH into the MPN facilitates sexual behavior and penile erection in male rats (68, 216, 464). In females, α-MSH stimulates sexual receptivity and lordosis, which is blocked in a dose-dependent manner by an MC3/4R antagonist (102, 182, 405). Drugs that target the melanocortin system have also been tested in humans as a treatment for male erectile dysfunction and female sexual dysfunction (83, 391, 485). A MSH analog, melanotan-II (MT-II), has been found to induce erections in men (485), and one study found that a melanocortin 3/4 receptor agonist, bremelanotide, increased satisfaction levels in men who were also taking Viagra (391). Increased understanding of the mechanisms underlying these actions may assist with sexual deficits in both men and women.

Recent work has begun to shed light on the neural pathways responsible for these effects. We recently examined the sexual behavior of mice expressing MC4R only on Sim1-positive neurons (found in the medial amygdala and PVH) in comparison with reactivable (tb)MC4R null mice. While tbMC4R null mice showed a longer latency to mount, a reduced intromission efficiency, and an inability to reach ejaculation, expression of MC4R only on Sim1 neurons reversed these sexual deficits. The amygdala is well known for its involvement in penile erection, sexual sensation, and orgasm in humans (176, 226, 438). Another possible location for the relevant POMC projections underlying these effects is the PVH. Viral tracing has established connections between the corpus cavernosus of rats and the PVH (295). In addition, lesions of PVH neurons delayed ejaculation and reduced intromission ratios (279). Oxytocin neurons in the PVH, besides having a critical role in lactation and parturition, also stimulate GnRH release (105, 163, 229, 375, 395). When injected centrally, oxytocin induces penile erection in rats, whereas the antagonist inhibits both sexual motivation and copulation (17, 303, 304). Another level of complexity and modulation of oxytocin in the PVH may be attained via projections from MC4R neurons in the PMV, a crucial olfactory relay in rodents (174). Establishing which pathway underlies the ability of melanocortins to improve sexual function has important clinical implications.

β-Endorphin also plays a critical, though opposing, role in the release of GnRH and sexual behavior. While α-MSH reduces food intake and stimulates sexual behavior and function, β-endorphin promotes food consumption and inhibits GnRH secretion (177, 182, 214, 266, 406). In line with these findings, the opioid antagonist naloxone increases LH secretion in mouse models and humans, probably due to blockade of β-endorphin inhibition of GnRH neurons (45, 82, 136, 177, 380). Opioid peptides inhibit the sexual behavior of females via projections to the VMHvl and downstream targets in the MPO and brainstem (353). This inhibitory role has also been observed in women. For example, opiate addicts often suffer from anorgasmia, amenorrhea, anovulation, and loss of libido. As in males, some of these effects result from the opioid inhibition of GnRH release, which leads to decreased production of gonadotropins and gonadal steroid hormones (41, 232, 360). Because of the dissociated effects of the two POMC neuropeptides (α-MSH and β-endorphin), investigators have proposed that a posttranslational mechanism or divergent neuronal circuits coordinate their physiological actions. In fact, electron microscopy has revealed that α-MSH and β-endorphin are located in separate pools of vesicles within the same neuron (250). These two pools of vesicles may be released with different time courses or by different regulatory stimuli (250).

Neurons expressing galanin-like peptide (GALP) comprise another ARH subpopulation associated with the metabolic control of reproduction. A negative energy balance decreases Galp expression. GALP administration induces LH secretion and restores the timing of pubertal maturation in calorie-restricted rats (154, 234, 265, 313). The effects of GALP seem to occur via direct actions on GnRH neurons (410). GALP administration also increases Kiss1 mRNA expression, but its effect on reproductive function does not require kisspeptin action (171). The metabolic cues associated with the control of GALP activity have not been clearly identified.

In addition to the ARH, the mediobasal hypothalamus consists of several nuclei housing first-order neurons involved in metabolic sensing. PMV neurons express receptors for leptin, insulin, ghrelin, and melanocortins (128, 174). They also project to brain areas associated with neuroendocrine and behavioral regulation (67, 174, 272, 386). These projections include direct innervation of GnRH and kisspeptin neurons and strong innervation of the VMHvl and MPO. The PMV has a crucial role as an integrative and modulatory site as it gathers signals from external environmental cues (i.e., odorants in rodents and photoperiod in seasonal breeders), reproductive status (sex steroids milieu), and energy availability (126, 251, 417). NO and glutamate are the main neurotransmitters, with the potential to stimulate downstream targets via glutamate and to coordinate adjacent neuronal activity via NO (34, 127, 129). Neurons in the MPO also express nNOS and respond to leptin. Because of the proximity to GnRH neurons and the volatile property of NO, the MPO nNOS neurons are well positioned to modulate the activity of GnRH neurons in response to leptin and, potentially, other metabolic cues (33).

Several other neuropeptides and hypothalamic sites may mediate the metabolic control of reproduction. For example, neurons expressing steroidogenic factor 1 (SF1), exclusively located in the VMH, play a key role in energy homeostasis (121, 242, 249). The lack of SF1 in the brain disrupts fertility, an effect that is likely due to developmental disruption of the VMH anatomical organization caused by this specific deletion (241). Melanin-concentrating hormone (MCH), orexin/hypocretin, and nesfatin in the lateral hypothalamic area play a role in energy balance and are also potential neuromodulators of the reproductive neuroendocrine axis (21, 169, 413, 422, 491, 500). However, the input and output of these neuropeptide systems and their interactions with metabolic cues have not been demonstrated. Similarly, RF-amide related peptide (RFRP, the ortholog of GnIH)-expressing neurons appear to play an important role, but the neuronal circuitry is still undefined (83, 84, 248, 377). Further studies are necessary to substantiate a position for these neuropeptides in the neuroanatomical framework of the metabolic control of reproductive function.

IV. METABOLIC DYSFUNCTION AND REPRODUCTIVE DEFICITS IN HUMANS

In addition to genetic mutations, metabolic dysfunctions such as eating disorders, common obesity, diabetes, and lipodystrophy in humans are associated with disruption of the reproductive axis, potentially via changes in the actions of metabolic signals in the HPG axis.

Approximately 20 million women and 10 million men in the United States suffer from anorexia nervosa, bulimia, binge-eating, or other unspecified feeding or eating disorders (423). Eating disorders in general are associated with amenorrhea, sexual dysfunction, infertility, miscarriage, and preterm delivery (14, 483). Pregnancy and childbirth rates are also reportedly lower in women with lifetime eating disorders compared with unaffected individuals (278). The menstrual irregularity observed in patients with anorexia nervosa likely results from suppression of GnRH release and impairments in the patterns of gonadotropin secretion (175). The suppressed secretion of LH and FSH leads to low levels of circulating estrogens and the cessation of ovulation. In adolescents with anorexia nervosa, the onset of secondary amenorrhea is usually observed after a 22% decrease in initial weight. In most cases, leptin levels predict amenorrhea and oligomenorrhea, and both are resolved when the body mass index (BMI) returns to 92.6% of the initial weight (461). Similarly, a general decrease in sexual libido and activity has been described in individuals with eating disorders (243, 358). The restoration of BMI in anorexic patients yields an increase in sexual drive, suggesting that the causal factor is hypothalamic hypogonadism secondary to energy insufficiency and decreased levels of leptin (243, 358).

Calorie restriction, which usually refers to a reduction in food intake without malnutrition, also impairs reproductive function in humans (331, 429). These deleterious effects vary according to the degree of caloric restriction. Mild and moderate reductions in calorie intake have no detrimental effect on fertility, whereas severe restriction suppresses reproductive function (296, 302, 507). A decrease in leptin levels seems to be the main signal for reproductive arrest, as leptin administration to women with hypothalamic amenorrhea due to extreme weight loss restores LH pulsatility and sex steroid levels (73, 276, 309, 476, 481).

On the other side of the spectrum, excess energy, as in obesity, also negatively affects human reproductive physiology. In girls, high adiposity is correlated with early puberty onset (42, 59, 204, 340, 472), which causes deleterious consequences for physical and psychological health. For example, earlier menarche and exposure to high levels of sex steroids increase the risk of adult obesity, T2D, and breast cancer (37, 115, 260, 350). Overweight and obese women have irregular menstrual cycles and ovulatory dysfunction. All these parameters are aggravated by a state of hyperleptinemia and leptin resistance. A blunted response to leptin leads to decreased activation of the HPG axis, low gonadotropin levels and impaired ovulation (3, 54, 307, 436).

Although obesity may cause reproductive deficits, high percentage of obese women are able to conceive. Obesity during pregnancy increases the risk of pregnancy complications, including maternal and fetal morbidity and mortality (70, 205). Obese women have higher risks of early and recurrent miscarriages, preterm births, stillbirths, and neonatal deaths (256, 372, 394). The effects of obesity on pregnancy complications are potentially associated with placental dysfunction not with HPG axis dysregulation (18, 349, 372, 394). In rodents, obesity induces placental necrosis, defective angiogenesis, and fetal resorption (287).

In obese men, testosterone is inversely correlated with fat mass and leptin due to a disruption of the HPG axis and increased aromatization of testosterone in adipocytes (223). Elevated body weight increases the risk of erectile dysfunction, potentially due to decreases in testosterone and disruptions of the melanocortin system, and impairs spermatogenesis and sperm motility (314, 457). However, obesity in men has also an effect on germ cells. Sperm from obese men have altered patterns of DNA methylation and higher rates of DNA fragmentation (103). In rodents, obesity decreases sperm motility and fertilization capacity (50). Following assisted reproductive technology, sperm from obese men resulted in deficient blastocyst development and decreased pregnancy rates (239).

In addition to having a direct impact on fertility, obesity often leads to high glucose levels, hyperinsulinemia, and diabetes. Both type 1 diabetes (T1D) and T2D have a negative impact on fertility (307). Up to 40% of females with T1D display menstrual disturbances (i.e., amenorrhea, oligomenorrhea, and menorrhagia), hyperandrogenism, and/or early menopause (480). In addition, late age at menarche is observed in girls with T1D (89, 162). The incidence of menstrual irregularities and menarche delays is higher in individuals with poor glucose control and higher BMI (162). Severe hypogonadism and low fertility rates have also been observed in T1D women (90). The few studies conducted in T1D men have revealed testosterone deficiencies in ~10% of these patients (78, 212), and its prevalence bears a direct relationship with waist circumference, triglyceride levels, and insulin requirements (78). It is possible that altered levels of glucose, FFA, and insulin disrupt the HPG axis in T1D patients. In addition, glucose metabolism in the testes, necessary for spermatogenesis and sperm quality and motility, may also be affected in T1D men (122).

Hypogonadotropic hypogonadism is the most common reproductive deficit of obesity and T2D (348). Individuals with T2D show relative decreases in LH pulsatility and disrupted sex steroid feedback, resulting in low gonadotropin levels (112). In women, T2D may induce early age at menopause (162, 480), and in men, low levels of testosterone may lead either directly or indirectly to erectile dysfunction (281). Men with erectile dysfunction are more likely to have undiagnosed diabetes (421). Whether this association results from decreased testosterone is not fully resolved, but studies in animal models have shown that neurite sprouting and length as well as the expression of growth factors are reduced in diabetic rats (298). It is also possible that in T2D and prediabetic states, insulin and leptin resistance in melanocortin neurons promotes erectile dysfunction by reducing α-MSH production, as observed in mice (207).

Insulin resistance and altered GLP1 secretion in obese and lean women have been associated with polycystic ovary syndrome (PCOS), one of the most common endocrine disorders, affecting 20% of women of reproductive age (132, 339, 366, 435, 469). Hyperinsulinemia increases LH secretion, cytochrome P-450 expression, and androgen production by the ovary as well as ACTH-mediated adrenal androgen production. High androgen levels in turn impair fertility by inducing follicle atresia and inhibiting meiotic maturation of germ cells (369), and increasing GnRH and LH pulsatility, potentially generating an uncontrolled feedforward cycle (58, 385). Metformin, an insulin-sensitizing medication, significantly reduces the production of hydroxyl-progesterone, testosterone, and androstenedione, improving the symptoms of PCOS (117).

PCOS is diagnosed by the presence of two of the following criteria: oligo or amenorrhea (anovulation), hyperandrogenemia (often visible as acne or hirsutism), and polycystic ovaries (117). In some cases, the ovaries develop “cysts” in the form of antral follicles that luteinize prematurely and fail to develop to a preovulatory stage. However, PCOS is a diagnosis of exclusion, and the development of a biochemical test of diagnostic criteria would be extremely useful in the clinic. The levels of anti-Mullerian hormone (221), insulin, or other circulating hormones have been suggested as potential biomarkers (475). Both genetic and environmental factors contribute to the development of PCOS, but the mechanisms are not entirely understood (117, 196). Hyperinsulinemia and hyperandrogenemia may both affect and disrupt the HPG axis.

Another metabolic disorder with a direct effect on reproductive function is lipodystrophy, characterized by the selective loss of adipose tissue. Lipodystrophy may be manifested in the whole body (generalized) or specific fat depots (partial) and may be genetic or acquired (277). The generalized form has been associated with mutations in 1-acylglycerol-3-phosphate O-acyltransferase (AGPAT2), seipin lipid droplet associated gene (BSCL2), caveolin 1 (CAV1), and polymerase I and transcript release factor (PTRF) genes. The partial form is caused by mutations in distinct genes, among which lamin A/C (LMNA) is the most common (277). Decreased fat depots result in low leptin levels and, as expected, most lipodystrophy features resemble a state of leptin deficiency. Increased food intake, ectopic fat storage, and insulin resistance are usually observed, ultimately leading to diabetes, nonalcoholic liver disease, PCOS, and infertility (192, 231, 455, 465). Leptin therapy improves insulin sensitivity, reverses PCOS, and restores fertility (282, 286).

V. FINAL REMARKS AND FUTURE CHALLENGES

The development of new techniques in gene editing and brain mapping have produced a considerable amount of data and insight on the integration of metabolism and reproduction by specific neuronal populations. In the last decade, metabolic cues with a role in reproductive function, including hormones, metabolites, and fuels, have been found to target specific neurons in the mediobasal hypothalamus, previously known to control energy balance and integrate environmental cues. These signals preferentially converge onto kisspeptin neurons upstream of GnRH neurons, expanding and refining the control of the reproductive axis. We have also learned that the selective actions of metabolic cues in subsets of neurons are sufficient to maintain fertility. However, none of these paths appears to be required for the maintenance of reproductive function, strengthening the case for circuit redundancy to ensure species survival. Studies using CRISPR/Cas9 technology or derivatives will help dissociate the heterogeneity of the system, e.g., specific kisspeptin, GABAergic, or nitrergic neuronal populations and their potential circuitry. Inducible gene mutations to bypass developmental compensations will also be necessary to challenge the system and dissociate congenital from acquired disorders. This approach will allow the interrogation of untested questions, such as the role of SF1 in the metabolic control of adult reproduction. Finally, the development of new technologies for brain imaging coupled with the remote activation/inhibition of specific pathways with the use of DREADDs and optogenetic technologies will advance the understanding of the network engaged in the control and integration of these complex systems crucial for individual and species survival.

Note ADDED IN PROOF

A paper published during proof revision demonstrated that the peptide PACAP (pituitary adenylate cyclase activating polypeptide) coexpressed in LepRb cells of the PMV also has a role in the metabolic control of reproduction (Ross RA, Leon S, Madara JC, Schafer D, Fergani C, Maguire CA, Verstegen AM, Brengle E, Kong D, Herbison AE, Kaiser UB, Lowell BB, Navarro VM. PACAP neurons in the ventral premammillary nucleus regulate reproductive function in the female mouse. eLife 7: e35960, 2018. doi: 10.7554/eLife.35960).

GRANTS

Studies in our laboratories are supported by National Institutes of Health Grants R01 HD069702 and R21 HD0900567 (to C. F. Elias) and R01 HD081792 (to J. W. Hill). We also thank the Michigan Diabetes Research Center (P30DK020572) for pilot grants (to C. F. Elias and J. W. Hill) and the Reproductive Sciences Program for pilot grants (to C. F. Elias).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.