1. Introduction

The gonadotropin-releasing hormone (GnRH) neurons represent the critical output cells of the neuronal network controlling fertility in all mammalian species [46] and [63]. These neurons have the unique property of migrating from the nose into the brain during embryogenesis [113], [88], [104] and [15] and exist, postnatally, as a scattered population of cell bodies within the basal forebrain [46]. Most GnRH neurons send axon projections to the median eminence of the hypothalamus from where they secrete GnRH into the pituitary portal system in an episodic manner to generate the pulsatile pattern of luteinizing hormone (LH) and follicle-stimulating hormone (FSH) required for fertility [46], [63] and [34]. The mechanisms through which the scattered GnRH neuron population generate an apparently synchronized output at the median eminence are not known and this remains a key issue of investigation within the field [73]. Various models have been proposed that involve roles for intrinsic episodic activity within the GnRH neurons themselves through to rhythm generation by the complex network of neurons regulating GnRH neuron activity [46] and [73].

One of the most intriguing aspects of GnRH neuron biology encountered to date has been their ability to generate spontaneous [Ca²⁺]_i transients. Because GnRH neurons exist in relatively low numbers and are diffusely distributed in the basal forebrain, several model systems have been developed to study the function of GnRH neurons. They comprise immortalized GnRH-secreting cell lines, explant cultures of embryonic GnRH-expressing cells derived from the nasal placode, and the acute brain slice preparation. Importantly, GnRH neurons in all of these model systems have been shown to exhibit spontaneous [Ca²⁺]_i transients. The purpose of this paper is to review the nature of spontaneous [Ca²⁺]_i transients in model GnRH neurons, including their mechanistic origins, regulation and modulation. In so doing we will discuss, initially, the results from the different model systems used to observe calcium dynamics in GnRH neurons, and then the evidence relating [Ca²⁺]_i dynamics to episodic hormone secretion. We will draw on data from the GnRH neuron field, as well as others, to bring together observations that may help explain the biological relevance of cytoplasmic [Ca²⁺]_i dynamics to GnRH neuron physiology and hormone secretion.

2. Immortalized GnRH-secreting cell lines

2.1. Calcium dynamics in immortalized GnRH-secreting cell lines

Genetically targeted tumorigenesis in the mouse has been used to generate immortalized cell lines that express GnRH. The most widely used of these cell lines has been the GT1 pedigree (GT1-1, GT1-3, GT1-7) generated by Mellon and colleagues [72]. GT1 cells express GnRH mRNA [72] and can secrete GnRH both in response to depolarization [72] and spontaneously, in an episodic manner, with a pulse interval of 25 min [58] and [70].

The sporadic nature of [Ca²⁺]_i transients in GT1 cells correlates well with electrophysiological findings in these cells that exhibit a variety of patterns of electrical activity including irregular recurrent short bursts of 4–5 action currents [16]. This correlation was subsequently confirmed by experiments in which calcium imaging was performed simultaneously with electrophysiological recordings, showing an almost complete synchronization between bursts of action currents and calcium transients, with each burst being followed by a calcium transient [27] and [108] (Fig. 3A). Further evidence that the generation of [Ca²⁺]_i transients is dependent on action potentials, derives from the observation that treatment with the voltage-sensitive Na⁺ channel blocker, tetrodotoxin (TTX), blocked both action potentials and [Ca²⁺]_i transients, but treatment with the L-type VGCC antagonist, nimodipine, only resulted in blockade of Ca²⁺ transients [27]. These studies indicate that [Ca²⁺]_i transients in GT1 cells result from their electrical firing patterns.

A key signaling mechanism regulating [Ca²⁺]_i transients in GT1 cells appears to be cAMP. Treatments that increase the levels of cAMP in GT1 cells stimulate GnRH secretion and increase the frequency of spontaneous [Ca²⁺]_i transients [18], [53] and [59] whereas lowering cAMP attenuates GnRH secretion and reduces spontaneous [Ca²⁺]_i transients [115]. As GT1 cells express cyclic nucleotide gated (CNG) non-selective cation channels [18] and [110], it seems that CNG channels mediate cAMP-dependent electrical excitation, which subsequently stimulates [Ca²⁺]_i transients and GnRH secretion (Fig. 2A). In further support of this scenario, siRNA-mediated down-regulation of CNG channels diminished both spontaneous and forskolin-stimulated GnRH secretion and [Ca²⁺]_i transients [9].

2.2. Neurotransmitter regulation of [Ca²⁺]_i transients in immortalized GnRH-secreting cell lines

Many different neurotransmitters have been found to modulate the secretion of GnRH from GT1 cells [108] and [112]. In terms of calcium, GABA has been found to exert dose-dependent increases in the frequency of [Ca²⁺]_i transients in GT1 cells [44] and [92]. This response occurs through the GABA_A receptor (Fig. 1A) and is partially abolished by TTX but completely prevented by either nifedipine [92] or the removal of extracellular Ca²⁺ [44].

Agonists for all of the ionotropic glutamate receptors are also found to evoke rapid, dose-dependent increases in [Ca²⁺]_i, that require extracellular Ca²⁺ and are sensitive to nifedipine [92]. Interestingly, the [Ca²⁺]_i increase in response to glutamate is partially blocked by the NMDA antagonist MK-801, the AMPA/kainate antagonist CNQX, or the two together [91], but a complete block could only be achieved with nifedipine [92], suggesting that NMDA- and AMPA-receptors have both direct and indirect control of calcium influx via VGCCs.

As GT1 cells exist in a culture environment of only GnRH-secreting cells it is not surprising to find that GnRH itself exerts complex and potent actions upon GT1 cells [108]. With respect to calcium, GnRH is found to initiate a biphasic response consisting of an acute increase in [Ca²⁺]_i followed by a prolonged elevated plateau [58] and [107]. Whereas the acute increase appears to originate from store mechanisms, the plateau is dependent upon extracellular calcium entry through an uncharacterized store depletion-activated channel [107].

2.3. Synchronization

Early studies showed that a remarkable feature of GT1 cells was their ability to generate episodic GnRH release even when separate culture dishes were perifused together [112], implying a role for a diffusible substance, perhaps GnRH itself, as a synchronizing influence. In addition to a role for diffusible substances, gap junctions among cells within an individual culture dish also appear to be an important means of synchronization between GT1 cells [71], [51], [109] and [12]. With respect to calcium, the nature of any synchronized [Ca²⁺]_i events in GT1 cells appears to depend upon the sub-type of GT1 cell used as Ca²⁺ transients move as “waves” through GT1-1 cultures, but appear as unsynchronized individual foci in GT1-7 cultures [17].

2.4. Summary for immortalized GnRH-secreting cell lines

To summarize, [Ca²⁺]_i transients in GT1 cells arise in response to spontaneous or driven electrical activity that activates L-type VGCCs to allow calcium entry from the extracellular space (Fig. 2A). The amino acid transmitters GABA and glutamate, acting through GABA_A and NMDA/AMPA/kainate receptors, depolarize GT1 cells and, thereby, activate the same pathway leading to [Ca²⁺]_i transients (Fig. 2A). In contrast, GnRH evokes [Ca²⁺]_i increases in GT1 cells by activating release from calcium store-dependent mechanisms. These stores are not, however, used to generate the on-going activity-dependent calcium transients observed in GT1 cells. While the data from GT1 cells are reasonably consistent given the different culture conditions and methods of calcium synchronization used by the different GT1 sub-clones, it remains difficult to establish the physiological relevance of GT1 calcium dynamics to GnRH secretion in vivo. One particular obstacle when measuring Ca²⁺ influx and GnRH release from GT1 cells, is that it is impossible to differentiate between events occurring at the GnRH nerve terminal and cell body.

3. Embryonic GnRH neurons in organotypic nasal placode cultures

Because the GnRH neuron population arises from the nasal placode, several independent groups have used this unique property to establish organotypic cultures and generate a source of embryonic GnRH neurons. While organotypic cultures of embryonic slices allow one to assess the cell migration within a native cellular environment [105], organotypic cultures of isolated nasal placodes provide a large number of GnRH neurons for studying their migration and cellular physiology (monkey [98], rat [28], mouse [35] and sheep [31]. In embryonic nasal explant cultures, the olfactory placode, as well as areas along the migratory path of GnRH neurons (monkey only), are dissected out and placed in culture. After 1 week of culture, some cells that migrate away from the main tissue mass exhibit GnRH peptide immunoreactivity, and represent post-mitotic GnRH neurons en route to the hypothalamus. Such placode cultures show GnRH release in a pulsatile manner, with a species-dependent periodicity that is consistent with in vivo measurements [99], [36], [30] and [23]. Since GnRH release requires calcium influx [100], the relationship between [Ca²⁺]_i and secretion has prompted investigators to look more closely at [Ca²⁺]_i dynamics in this model where, additionally, many embryonic GnRH neurons can be monitored simultaneously.

3.1. Calcium dynamics in embryonic GnRH neurons

Many neuronal cell types exhibit [Ca²⁺]_i transients during embryogenesis [93] and, as such, it has not been surprising to find that embryonic GnRH neurons in all nasal placode-derived preparations examined to date exhibit robust on-going [Ca²⁺]_i transients (monkey [100]; mouse [23] and [75]). While a degree of variability is observed in the calcium dynamics displayed by individual cells within a culture, much larger differences are found across species in both the frequency (Fig. 1B) and the duration (Fig. 1B) of [Ca²⁺]_i transients; in the monkey, [Ca²⁺]_i transients occur in GnRH neurons every 8 min and last for 1.5 min [100] (Fig. 1B) whereas they occur much more frequently every 40 sec and last for only 12 sec [24] (Fig. 1B) in the mouse. In both preparations, the frequency of [Ca²⁺]_i transients appears to increase with time in culture [100], [75] and [25].

The mechanisms underlying spontaneous [Ca²⁺]_i transients in embryonic GnRH neurons has been investigated in both the mouse and monkey models. In the monkey, TTX was reported to suppress the firing of embryonic GnRH neurons [1], but have no effect upon [Ca²⁺]_i transients [2], suggesting that [Ca²⁺]_i transients occur independently of electrical activity. This possibility received indirect support from the finding that three categories of GnRH neurons could be described based on the temporal characteristics of individual [Ca²⁺]_i transients [100], but similar categories were not evident based on electrical activity [1]. However, since both [Ca²⁺]_i transients and burst firing occurred on 1 min time-scale, a relationship between [Ca²⁺]_i and electrical activity in monkey embryonic GnRH neurons system cannot be ruled out at present.

For mouse embryonic GnRH neurons, studies have shown both complete [24] and partial [75] and [97] dependence of [Ca²⁺]_i transients upon TTX/electrical activity. This variable TTX dependence may, however, be dependent upon the age of the culture as [Ca²⁺]_i transients have recently been found to be abolished completely by TTX in young (7 days in vitro) but not old (21 days in vitro) mouse placode GnRH neurons [25]. Importantly, dual [Ca²⁺]_i -electrical recordings from single embryonic GnRH neurons at 6–10 days in vitro indicate that [Ca²⁺]_i transients are well-correlated with bursting electrical activity [22] (Fig. 3B). Thus, it appears that electrical activity can underlie [Ca²⁺]_i transients in embryonic GnRH neurons, but that other developmentally-regulated mechanisms are also likely to be involved in generating [Ca²⁺]_i transients in the placodal culture preparation (Fig. 2B and C).

The channels and pathways involved in generating [Ca²⁺]_i transients in embryonic GnRH neurons have been examined only in the mouse placode cultures. Single cell RT-PCR analysis indicated that GnRH neurons express both N- and L- type VGCC [103] and functional evidence for high threshold VGCC, and more specifically L-type VGCC, has been provided by whole cell electrophysiological recordings [60]. The involvement of VGCC in the generation of [Ca²⁺]_i transients has been investigated using cadmium, a broad spectrum calcium channel antagonist. Although 40% of the embryonic GnRH cells showed a decrease in the frequency of the [Ca²⁺]_i transients with cadmium, only 15% showed a complete arrest [24]. The role of internal calcium stores in the generation of [Ca²⁺]_i transients has also been examined. The depletion of ionomycin-sensitive calcium stores was found to have no effect on [Ca²⁺]_i transients, and only a minority of cells showed a reduced frequency of transients in response to the blockers of endoplasmic reticulum calcium channels, ryanodine receptors (RyR) and inositol 1,3,4-triphosphate receptors (IP₃R) [23]. Together, these studies indicate that internal stores are not involved in the generation of [Ca²⁺]_i transients in mouse embryonic GnRH neurons and that, while some role appears to exist for electrical activity and VGCC, another unknown pathway is also involved (Fig. 2C).

3.2. Amino acid neurotransmitter regulation of [Ca²⁺]_i transients in embryonic GnRH neurons

In mouse nasal placode-derived cultures, GABA has been identified as a neurotransmitter eliciting [Ca²⁺]_i transients via the activation of the GABA_A receptor [24], [75], [21] and [22]. These data are consistent with the depolarization and GABA-induced action potential firing observed electrophysiologically in embryonic mouse GnRH neurons [60]. Acute treatment with the GABA_A receptor agonist muscimol induced an initial rapid increase in [Ca²⁺]_i followed by a sustained plateau (Fig. 1B). Whereas the initial rise was insensitive to TTX, the plateau was dependent upon voltage-gated sodium channels [74], raising the possibility that GABA_A receptor activation in some way initiates release from internal calcium stores followed by more classical VGCC-dependent calcium flux through the membrane.

Similar to GABA, glutamate has also been shown to evoke action potential firing in mouse embryonic GnRH neurons [60] and recent data indicate that endogenously activated non-NMDA glutamatergic receptors can trigger [Ca²⁺]_i transients in these cells [25]. Importantly, recent work has also shown that [Ca²⁺]_i transients in embryonic GnRH neurons cultured for 7 days are dependent upon GABA and glutamate receptor activation, as treatment with GABA_A- and AMPA-receptor antagonists largely abolishes [Ca²⁺]_i transients.

3.3. ATP and CCK regulation of [Ca²⁺]_i transients in embryonic GnRH neurons

In monkey nasal placode-derived cultures, [Ca²⁺]_i changes were observed in GnRH and non-GnRH neurons as well as in non-neuronal cells, suggesting the possible contribution of other cell types to the genesis of the [Ca²⁺]_i transients in GnRH neurons [100]. Hypothesizing that ATP could be the chemical substrate promoting calcium waves across the placodal cultures [101], Terasawa and co-workers have examined the calcium response of GnRH neurons to ATP. This showed that exogenous ATP could evoke a P2X receptor-dependent calcium response from embryonic GnRH neurons and that apyrase, an enzyme degrading ATP, arrested [Ca²⁺]_i transients in these cells. This suggests an important role for ATP in the generation of on-going [Ca²⁺]_i transients in monkey placode GnRH neurons.

In mouse nasal placode-derived cultures, cholecystokinin (CCK) has recently been shown to modulate spontaneous [Ca²⁺]_i transients in GnRH neurons via activation of CCK-R1 receptors [37]. In contrast with “young” placodal cultures during which GABA and glutamate underlie most of the [Ca²⁺]_i transients, a recent study identified CCK as a third input involved in the endogenous triggering of calcium oscillations in older cultures [25]. The specific signaling pathway was not investigated but both protein kinase A- and C-dependent phosphorylation, triggered by G-protein coupled receptor activation, alter GnRH neuronal activity [21] and [24] and thus may be involved.

3.4. Synchronization

The potential synchronicity of the GnRH cell population has been assessed using calcium imaging techniques in both monkey and mouse placode models. In both preparations, some [Ca²⁺]_i transients appear to occur simultaneously across different cells within the same culture [100] and [75] and, intriguingly, these episodes of synchronized transients occur with a similar periodicity to GnRH pulse frequency in vitro (monkey, [99] or in vivo (mouse, [38]. In mouse, the simultaneous recording of [Ca²⁺]_i transients and GnRH release showed that at least 30% of the recorded cells displayed a [Ca²⁺]_i transient prior to a GnRH pulse [23], suggesting a correlation between synchronized calcium events and episodic secretion.

The cellular mechanisms that underlie this occasional synchronization have been examined in both mouse and monkey placode preparations. In the embryonic mouse, gap junctions do not seem to be important as no evidence was found for dye coupling between individual GnRH neurons [60]. A role for classical synapses and/or action potential-dependent release in mediating this co-ordination also seems unlikely as synchronized [Ca²⁺]_i transients persist in embryonic GnRH neurons in the presence of TTX, in both monkeys and mice [75] and [2]. These studies would not, however, rule out a role for molecules such as ATP, GABA and glutamate that can be released in an action potential-independent manner. Indeed, the importance of ATP for [Ca²⁺]_i transients, discovered in the monkey placode cultures [101], supports this role as does the key role that endogenous GABA plays in modulating the occurrence of synchronized [Ca²⁺]_i events in the mouse [75]. Interestingly, the application of 17-β-estradiol in both monkey and mouse models facilitates synchronicity [2] and [97] although the physiological impact of this estradiol treatment remains unknown [47].

Despite these insights, the mechanism through which multiple cells achieve occasional synchronous [Ca²⁺]_i transients remains unknown. It is important to note that every factor influencing synchronicity also modulates the frequency of [Ca²⁺]_i transients in individual cells, i.e. the proportion of synchronized events is positively correlated with the frequency of [Ca²⁺]_i transients [75] and [2]. Therefore, it is hard to exclude the possibility that synchronous events are simply determined by the frequency of [Ca²⁺]_i transients – i.e. the probability of synchronization increases as the individual frequency increases.

3.5. Summary embryonic GnRH neurons

To summarize, [Ca²⁺]_i transients are observed in embryonic GnRH neurons derived from both the monkey and mouse nasal placode, although their temporal dynamics are quite different (Fig. 1B). Despite fairly intensive investigation, the mechanisms underlying [Ca²⁺]_i transients in embryonic GnRH neurons remain obscure. It would seem that internal store mechanisms are not involved and, while electrical activity and VGCC may be responsible for some calcium events, another uncharacterized mechanism/pathway involving extracellular calcium entry is very likely to be present (Fig. 2B and C). There is now emerging evidence in the mouse that embryonic GnRH neurons in young cultures are dependent entirely upon on-going GABA and glutamate signaling to generate [Ca²⁺]_i transients while older cultures develop additional mechanisms of control. Calcium transients in embryonic GnRH neurons can be driven by amino acids, neuropeptides and ATP at all ages and this may be relevant to the mechanism underlying the occasional synchronization of individual [Ca²⁺]_i transients observed in placodal cultures. As with GT1 cell preparations, the inability to differentially modulate calcium at the terminal and cell body in culture, means that the relationship between cell body [Ca²⁺]_i transients and episodic GnRH release from the nerve terminal is unknown.

4. Adult GnRH neurons in acute brain slices

Since the development of GnRH-reporter transgenics [48], the acute brain slice preparation has been used extensively to study the physiology of postnatal GnRH neurons. This system has the advantage of facilitating examination of adult GnRH neurons in their native environment, although, it is important to note that their distant synaptic inputs are severed. Recently, investigators have created a line of transgenic mice in which the GnRH promoter is used to drive expression of a genetically encoded calcium indicator (GECI), Ratiometric Pericam [76], specifically in GnRH neurons [52]. By making acute brain slices from GnRH-Pericam transgenic mice, the [Ca²⁺]_i dynamics of soma and proximal dendrites of living adult GnRH neurons can be recorded in real time.

GECIs differ in a number of ways from the synthetic calcium indicators that have been used to report [Ca²⁺]_i in in vitro GnRH model systems. Synthetic indicators have high-affinity for calcium (10.4 μM for Fluo4-FF [114]), wide dynamic range [83], and fast kinetics making them highly sensitive to rapid changes in [Ca²⁺]_i even if such changes are relatively small, e.g. associated with single action potentials (APs). By contrast, GECIs typically have lower affinity for calcium (e.g. 1.7 μM for Ratiometric Pericam [76], become saturated within a narrower concentration range [83], and consequently can have reduced sensitivity to small and/or rapid changes in [Ca²⁺]_i [83]. Thus, GECIs are very sensitive to reporting high, longer duration calcium events [83] but less able to detect very fast events such as those associated with an individual action potential. At present it is only possible to record [Ca²⁺]_i in GnRH-Pericam neurons every 0.5 s.

4.1. Calcium dynamics in adult GnRH neurons in mouse brain slices

The majority (70%) of adult GnRH neurons have been found to exhibit [Ca²⁺]_i transients that last for 10 s and occur over a range of frequencies but with a mean of 7 per hour [52] (Fig. 1C). The dynamics of [Ca²⁺]_i transients appear very similar in GnRH neurons from adult males and females, with younger ages yet to be examined.

Simultaneous recordings of calcium dynamics and firing in individual adult mouse GnRH neurons have revealed that [Ca²⁺]_i transients exist only in GnRH neurons that fire in a bursting pattern and that these transients are correlated perfectly with each burst of action potentials [61] (Fig. 3C). As with the other models using different calcium indicators, changes in [Ca²⁺]_i are only rarely detected associated with single action potentials. Depolarizing stimuli that generate burst firing in GnRH neurons were found to evoke [Ca²⁺]_i transients, and TTX abolished both firing and [Ca²⁺]_i transients [61]. These observations demonstrate that action potential firing initiates [Ca²⁺]_i transients in adult GnRH neurons.

As noted above for GT1 and embryonic GnRH neurons, VGCC play a key role in the generation of [Ca²⁺]_i transients. Present data indicate that adult rodent GnRH neurons express all major sub-types of the VGCC family. Calcium currents recorded from either dissociated primary GnRH neurons or GnRH neurons in the slice preparation have found evidence for L, N, P/Q, R and T-type channels [89], [55] M. Kato, K. Ui-Tei, M. Watanabe and Y. Sakuma, Characterization of voltage-gated calcium currents in gonadotropin-releasing hormone neurons tagged with green fluorescent protein in rats, Endocrinology 144 (2003), pp. 5118–5125. Full Text via CrossRef | View Record in Scopus | Cited By in Scopus (43)[55], [78], [90] and [116]. In addition, RT-PCR has demonstrated the presence of T-type channel subunit mRNAs in mouse GnRH neurons [116]. As [Ca²⁺]_i transients in adult GnRH neurons are abolished by both calcium-free bathing medium and nifedipine, it would appear that the L-type VGCC is the key sub-type underlying plasma membrane calcium entry [61]. It is also clear, however, that internal calcium store mechanisms play an important role in the generation of [Ca²⁺]_i transients as both IP₃R and SERCA pump inhibitors abolish [Ca²⁺]_i transients in adult GnRH neurons [52] and [61].

Together, these data indicate that [Ca²⁺]_i transients arise in adult GnRH neurons as a result of burst firing that causes an influx of extracellular calcium through L-type VGCCs that is then amplified by an IP₃R-dependent calcium-induced calcium release store mechanism to generate the 10 s-duration [Ca²⁺]_i transient (Fig. 2D).

4.2. Neurotransmitter regulation of [Ca²⁺]_i transients in adult GnRH neurons in situ

Both GABA and glutamate are powerful regulators of GnRH neuron electrical activity, making them prime candidates for modulating [Ca²⁺]_i transients. Using GnRH-Pericam transgenic mice, treatment with GABA, NMDA or AMPA was found to elicit an immediate rise followed by a decline in [Ca²⁺]_i in adult GnRH neurons, even in the presence of TTX. The GABA effect was reproduced by muscimol, a selective GABA_A receptor agonist (Fig. 1C) but not by baclofen, a selective GABA_B receptor agonist [26]. Approximately 65–70% of adult GnRH neurons responded to either AMPA or GABA in both male and female mice. By contrast, only a few (20%) adult GnRH neurons responded to NMDA regardless of sex. The mechanisms underlying the ability of GABA and AMPA to stimulate an increase in [Ca²⁺]_i in adult GnRH neurons were different; for GABA, the initial increase in [Ca²⁺]_i originated from GABA_A receptor-mediated activation of L-type VGCCs, whereas for AMPA this appeared to involve direct calcium entry through the AMPA receptor. This initial increase in [Ca²⁺]_i then evoked the same calcium-induced calcium release from IP3R-dependent store mechanisms for both GABA and AMPA (Fig. 2D). These observations show that all of the principal amino acid receptors are able to modulate [Ca²⁺]_i in adult GnRH neurons, but in an intracellular pathway-specific manner [26].

In contrast to GABA and glutamate, GnRH peptide was found to have no abrupt effect upon [Ca²⁺]_i in adult GnRH neurons; in the presence of TTX, GnRH generated a small but significant gradual decline in [Ca²⁺]_i in the majority of GnRH neurons in the adult female mouse [45].

4.3. Synchronization

While the GnRH-Pericam mouse has enabled single GnRH neurons to be examined in real time, it is rare to get several cells in the same plane of focus in the same field of view. Thus, only very limited information can be gained on the potential co-ordination of [Ca²⁺]_i levels between adult GnRH neurons in situ. At present, no synchronization of [Ca²⁺]_i transients has been observed in up to 5 simultaneously recorded GnRH neurons in the coronal brain slice (Romano and Herbison, unpublished observations). Although there has been some recent insight into the potential synchronization mechanisms used by adult GnRH neurons [14] this issue remains poorly understood.

4.4. Summary of adult GnRH neurons

Much less work has been undertaken on the calcium dynamics of adult GnRH neurons compared to embryonic or GT1 cells. However, information gathered to date shows that [Ca²⁺]_i transients are an exclusive feature of GnRH neurons that exhibit burst firing. The [Ca²⁺]_i transients observed in these cells arise from action potential activation of L-type VGCCs allowing small amounts of plasma membrane calcium flux that are then amplified by IP₃R-dependent store mechanisms (Fig. 2D). While this mechanism accounts for the “spontaneous” [Ca²⁺]_i transients observed in adult GnRH neurons, it also underlies the ability of neurotransmitters like GABA and glutamate to generate acute changes in [Ca²⁺]_i. Although it may be inferred that action potential firing represents release of GnRH at the median eminence, it remains that the relationship between the pattern of firing displayed by an individual GnRH neuron and its effect on GnRH secretion is unknown.

5. Common features of calcium dynamics in GnRH neurons

It is clear from all studies undertaken so far that cell body [Ca²⁺]_i transients are an integral feature of GnRH neurons regardless of developmental age (Fig. 1). Although it was not surprising to find [Ca²⁺]_i transients in embryonic GnRH neurons, the observation of long-duration (i.e. 10 s) [Ca²⁺]_i transients in adult GnRH neurons is unique in the mature nervous system [52]. Importantly, the dynamics of [Ca²⁺]_i transients appear to be remarkably similar in mouse GT1, embryonic and adult GnRH neurons (Fig. 1). This indicates not only similarity in mechanism, and thus biological reality across model systems, but also that technical variability from the use of different calcium indicators is not likely to be a significant concern for comparing data from different systems. For example, the major difference in [Ca²⁺]_i transient dynamics is found between monkey (90 s-duration) and mouse (10 s-duration) models despite the same calcium indicator having been used. In the mouse, all three model systems (GT1, young placodal and adult GnRH neurons) exhibit [Ca²⁺]_i transients driven by burst firing (Fig. 3). Thus, it would appear as though burst firing and the consequential [Ca²⁺]_i transients are a signature feature of the GnRH neuron throughout development into adulthood, for which there may be a common purpose. Interestingly, studies undertaken in the monkey have shown the same general phenomenon, although the size and frequency of [Ca²⁺]_i transients are remarkably different and evidence that they originate from cell firing is presently lacking. Hence, it is not yet clear whether the monkey [Ca²⁺]_i transients are related to those found in mice.

Although it appears that the burst firing-[Ca²⁺]_i transient “signature” is common to mouse GnRH neurons at all developmental stages, the mechanisms underlying the generation of the [Ca²⁺]_i transients appear to vary with developmental age. In all preparations (GT1, young embryonic and adult), the [Ca²⁺]_i transients are dependent upon cell firing and L-type VGCC. However, whereas there is no role for intracellular calcium stores in GT1 and embryonic GnRH neurons, IP₃R-dependent calcium-induced calcium release from stores is essential for [Ca²⁺]_i transients in adult GnRH neurons (Fig. 2). Indeed, the involvement of stores explains how adult GnRH neurons can achieve relatively prolonged [Ca²⁺]_i transients. However, since placodal GnRH neurons also have prolonged [Ca²⁺]_i transients, but without evidence of store involvement, there is a need to define the role of other factors such as calcium buffering, transport, and the role of differential synaptic inputs between the two model systems.

Ultimately, the occurrence of [Ca²⁺]_i transients is dependent upon the, as yet uncertain, mechanics behind burst firing in GnRH neurons [46] and [73]. As expected, key neurotransmitters such as GABA and glutamate, that are potent regulators of GnRH neuron firing, have also been found to exert acute effects upon [Ca²⁺]_i in all preparations of GnRH neurons examined to date. In some cases, such as young embryonic GnRH neurons, these amino acid neurotransmitters have even been found to be critical for [Ca²⁺]_i transients to occur. In contrast to GABA and glutamate, there has been much less consistency reported for the effects of other neurotransmitters, several of which have only been examined in one type of preparation. Of note, however, there are very marked differences in the effects of GnRH on GT1 cells and adult GnRH neurons; in the former GnRH generates an acute biphasic stimulatory response whereas it generates a slow inhibitory effect in the latter. This may be related to the special significance of GnRH as a neurotransmitter in the GT1 cultures where there are only GnRH-secreting cells compared with the brain where GnRH inputs to GnRH neurons are rare compared with GABA and glutamate signaling [46].

6. Biological significance of [Ca²⁺]_i signaling in GnRH neurons

A clear and possibly distinctive feature of GnRH neurons is their ability to exhibit activity-dependent, long-duration [Ca²⁺]_i transients. The physiological roles of these transients presently remain unclear with a vast array of possibilities [6] and [7]. It is important at the outset to be clear that these long-duration [Ca²⁺]_i transients have only been recorded from the GnRH neuron cell body and that the dynamics of [Ca²⁺]_i events at the GnRH neuron nerve terminal remain unknown. Nevertheless, it is rather likely that [Ca²⁺]_i dynamics within the GnRH nerve terminal are typical of any presynaptic nerve terminal operating classical calcium-dependent vesicular release mechanisms [84] and [4]. Given the large distances between terminal and cell body in situ, it is unlikely that changes in [Ca²⁺]_i recorded from GnRH neuron cell bodies and proximal dendrites have any direct role in regulating GnRH secretion from the terminal. Moreover, in other neurosecretory cells, the complement of ion channels in the cell body is different from that in the nerve terminal [29], further indicating fundamentally distinct functions for [Ca²⁺]_i dynamics at each cellular location. Finally, peptide secretion can be induced by stimulation of the terminals directly, and in complete isolation from the cell body [84] and [20], indicating that cell body activity is not absolutely required for peptide secretion, although there may be a modulatory relationship. Thus, the influence of cell body/dendrite [Ca²⁺]_i changes on GnRH secretion would be expected to occur indirectly through the modulation of ion channels on the dendrite and soma that determine the generation of action potentials in GnRH neurons.

6.1. Calcium regulation of gene transcription in GnRH neurons

In addition to its well-known role as a facilitator of secretion, calcium influx is now appreciated to alter neuronal physiology in a number of ways, including the activation of context-specific programs of gene expression [33], [43] and [118]. Since the discovery that c-Fos is an activity- and calcium-dependent transcription factor [42], a staggering number of transcriptional modulators have been reported to be controlled by changes in [Ca²⁺]_i in a number of cell types [77], [19] and [65]. In addition, it appears that the route of calcium entry is important in defining the spectrum of gene expression changes, and is likely due to differences in the temporal and spatial characteristics of [Ca²⁺]_i elevation [43] and [39]. More recently, changes in [Ca²⁺]_i in the nucleus have been shown to be critical in controlling the timing of turning gene expression off and on [80] and [66], and adding a further dimension to calcium signaling. Together these data suggest a scenario in which the intrinsically and extrinsically determined [Ca²⁺]_i dynamics of a given cell are coupled with its unique repertoire of transcriptional modulators in order to give rise to a context-relevant change in gene expression that underlies its physiology.

For cells involved in learning, [Ca²⁺]_i changes result in patterns of gene expression that faciliate and consolidate specific synapses (see for example [50]). For cells whose main business is secretion, [Ca²⁺]_i dynamics are anticipated to regulate gene expression associated with this function. Thus, GnRH neurons are likely to undergo substantial physiological changes in preparation for secretion, including increased expression of genes encoding GnRH or GnRH peptide processing proteins. Indeed, there is evidence in GT1 cells for episodic GnRH gene expression [79] that may represent such a preparatory event. In addition, [Ca²⁺]_i transients in GnRH neurons may affect hormone release indirectly by signaling biochemical changes that either drive or are prerequisite for secretion. For example, the calcium-regulated transcription factors CREB and c-Fos are well-described markers of GnRH neuron activation at the time of the preovulatory GnRH surge [49] and [3]. Although their genomic targets are currently unknown in GnRH neurons, these transcription factors likely serve as a key link between [Ca²⁺]_i dynamics and changes in gene expression that are critical to normal GnRH secretion. Nevertheless, although many exciting possibilities exist, the relationship between [Ca²⁺]_i transients, activation of downstream signaling molecules, and ultimately hormone secretion, still awaits experimental assessment.

6.2. Calcium transients help shape burst firing dynamics in adult GnRH neurons

A recent study has shown that one role of [Ca²⁺]_i transients in GnRH neurons is to shape burst firing in these cells [61]. Prior investigations in both placodal and acute brain slice models have reported that individual GnRH neurons can exhibit a range of different firing patterns [1], [21], [60] and [78]. The most distinctive and possibly common of these is the burst firing pattern, with 2–10 action potentials occurring as a cluster typically every 20 s [1], [61] and [78] (Fig. 3). Using dual electrical-calcium recording approaches, Lee and colleagues were able to show that the [Ca²⁺]_i transient in GnRH neurons regulates two calcium-activated potassium channels; one that controls principally the frequency and pattern of firing within the burst, and another that helps control the time interval between bursts [61] (Fig. 4).

The immediate rising phase of the [Ca²⁺]_i transient, that occurs at the beginning of a burst of action potentials in GnRH neurons, activates a small conductance calcium-activated potassium (SK) channel that then slows and eventually stops the burst, while the latter part of the [Ca²⁺]_i transient activates a very slow, UCL2077-sensitive calcium-activated potassium channel that acts to determine when the next burst will occur (Fig. 4). The SK channel is often encountered in mammalian neurons where it activates immediately after VGCC calcium entry to generate the after-hyperpolarization (AHP) and decays over 100 ms to control firing excitability [94]. Despite GnRH neurons exhibiting all of the normal SK subunit mRNAs [11] and [56], they express SK channels with a unique two phase decay, the normal fast decay and also a very slow (20 s) decay so far observed only in GnRH neurons [90], [56] and [68]. Many neuronal cell types also exhibit a slow calcium-activated potassium channel, that remains molecularly uncharacterized [87], and the GnRH neurons exhibit a UCL2077-sensitive version of this channel with a decay of 20 s that also contributes to their prolonged AHP [61]. Hence, GnRH neurons appear special in their use of prolonged [Ca²⁺]_i transients and slowly decaying calcium-activated potassium channels to control their burst firing. Together, these features result in a cycle of activity whereby a burst of action potentials generates a [Ca²⁺]_i transient that limits burst size as well as the time to the next burst, which then activates another [Ca²⁺]_i transient and so on. Interestingly, striatal cholinergic neurons have recently been shown to exhibit somewhat similar behaviors but on a different temporal scale [40], and possibly the closest representation of the scenario found in the GnRH neurons occurs in retinal amacrine cells where calcium spikes also regulate prolonged AHPs [117].

Although these studies have provided an insight into one of the roles of [Ca²⁺]_i transients in GnRH neurons, important unresolved issues reside around the functional significance of burst firing in GnRH neurons. The roles of calcium in regulating ion channels are quite different between GnRH and magnocellular neurons [64], [102] and [85], but burst firing is known to enhance the efficiency of hormone release in both vasopressin neurons [32] and [8] and oxytocin neurons [111] and [67]. On this basis, it is suspected that the bursting behavior of GnRH neurons is also important for pulsatile GnRH secretion [46], [73] and [62]. However, the relationship between GnRH neuron burst firing (every 20 s) and pulsatile GnRH secretion (every 30–60 min) is completely unknown and remains one of the key unanswered questions in the field of reproductive neurobiology.

In summary, through reviewing existing data on [Ca²⁺]_i dynamics in the different GnRH neuron model systems, it is apparent that most GnRH neurons generate [Ca²⁺]_i transients throughout development and into adulthood. In most cases, there is clear evidence that these transients are initiated by the burst firing patterns of GnRH neurons. For monkey and older mouse embryonic cultures that appear to exhibit [Ca²⁺]_i transients independent of their electrical activity, definitive dual calcium-electrical recording studies have yet to be undertaken. Review of the data also suggest that there are likely to be multiple different sub-cellular mechanisms through which a GnRH neuron can generate a [Ca²⁺]_i transient. By the time of adulthood, the mechanism appears to have matured into one in which both plasma membrane calcium flux and internal store components work interactively to establish a [Ca²⁺]_i transient of 10 s-duration. This very long transient represents a unique phenomenon within the mature nervous system. While there is much work yet to be undertaken before the full biological significance of the [Ca²⁺]_i transient is understood in GnRH neurons, one clear and important role is to regulate their patterns of burst firing. Further challenges also lie ahead in deciphering the roles of [Ca²⁺]_i within compartments of the GnRH neuron such as their extensive dendritic tree, spines and nerve terminals.