INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

IN GOLDFISH PITUITARY, growth hormone (GH) secretion is under the stimulatory influence of multiple neuroendocrine factors, including two endogenous gonadotropin-releasing hormone peptides (GnRH) and dopamine (DA). It has been established that GnRH and DA D₁ stimulation of GH release is mediated by protein kinase C- and cAMP/protein kinase A (PKA)-dependent mechanisms, respectively (61-63). In addition, the availability of extracellular calcium is essential for maintenance of these hormone-release responses. In particular, PKA phosphorylation of high voltage-activated Ca²⁺ channels (HVCCs) has been proposed to be a key component of the DA response. However, until recently, very little was known about the role that intracellular stores play in mediating agonist-evoked secretion in goldfish somatotropes. Recent studies (28) identified a caffeine-activated Ca²⁺ store that participates in GnRH-evoked release in goldfish somatotropes.

One of the major intracellular signaling cascades activated by caffeine is calcium-induced Ca²⁺ release (CICR), which is used to amplify and propagate Ca²⁺ signals throughout the cell to various target proteins (reviewed in Ref. 9). CICR is mediated through a family of intracellular channels called ryanodine receptors located on the sarco-/endoplasmic reticulum (reviewed in Refs. 43, 54, 64). The sensitivity of these receptors for activation by Ca²⁺ increases dramatically in the presence of caffeine, which allows channel opening at intracellular Ca²⁺ concentrations ([Ca²⁺]_i) that are present in resting cells (12). Moreover, these channels are classically modulated by ryanodine.

Originally characterized as an important Ca²⁺-releasing mechanism in muscle cells (reviewed in Refs. 12, 64), it is increasingly clear that CICR plays a critical role in many excitable (reviewed in Ref. 9) and nonexcitable cells (33). In rat somatotropes, a caffeine-/ryanodine-sensitive Ca²⁺ store affects GH-releasing hexapeptide- and GH-releasing hormone-evoked Ca²⁺ transients (21, 44) and secretion (23). Moreover, Ca²⁺ influx through voltage-gated Ca²⁺ channels appears to be a critical mechanism of regulating evoked GH secretion in rats (24, 36-38, 46) and may also be tightly linked to CICR (44; although see also Ref. 57). Similar caffeine-/ryanodine-sensitive mechanisms are also present in sympathetic neurons (16, 17, 35, 40, 51) and adrenal chromaffin cells (2, 20, 55). In contrast, in rat gonadotropes (30, 31, 52, 58, 59; reviewed in Ref. 53) and both rat and frog melanotropes (18, 34), the primary intracellular channels that mediate secretion as well as intracellular Ca²⁺ transients are a class of channels that are activated by inositol trisphosphate (IP₃) with little if any contribution from ryanodine-sensitive channels.

Although an extensive amount is known about intracellular signaling for evoked secretion in goldfish somatotropes, particularly with regards to cAMP and Ca²⁺ entry (reviewed in Refs. 7, 8), very little is known about how these aspects of signaling interact with the recently identified Ca²⁺ store(s). Understanding the nature of caffeine action and properties of the caffeine-sensitive Ca²⁺ store(s) is an important step in elucidating how caffeine-sensitive events mediate the physiological regulation of secretory activity in goldfish somatotropes, in particular, how the properties of secretion mediated by stores might differ from properties of secretion mediated by activation of Ca²⁺ entry. In the present study, we further characterize the GH response to caffeine as being partially sensitive to ryanodine and show that caffeine-evoked secretion is initiated independent of Ca²⁺ entry through Cd²⁺-sensitive HVCCs yet requires HVCCs for refilling of stores to generate a maintained GH secretory response. These mechanisms of action, particularly with regard to the initiation of the secretory response, are quite different from those mediating DA-evoked release.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Animals and cell preparation. Common goldfish (Carrasius auratus) of 10-13 cm in body length were purchased from Grassyforks Fisheries (Martinsville, IN) or Aquatic Imports (Calgary, AB, Canada) and kept in flow-through aquariums at 18°C on a natural photoperiod adjusted to the local Edmonton area. Fish were fed commercial fish food ad libitum. All animal-use protocols were approved by the University of Alberta Bioscience Animal Care Committee.

Solutions. Cell culture media were used as described previously (5). All media contained M-199 (GIBCO, Grand Island, NY) with L-glutamine, 26 mM NaHCO₃, 25 mM HEPES, streptomycin (100 mg/l), and penicillin (100,000 U/l) at pH 7.2 (adjusted with 10 N NaOH). Dispersion medium contained M-199 with Hanks' salts and 3 g/l BSA. Plating medium (for overnight incubation) contained M-199 with Earle's salts and was supplemented with 1% horse serum (GIBCO). Testing medium was the same as dispersion medium except that BSA was reduced to 1 g/l. Ca²⁺-free testing medium was the same as testing medium with the omission of CaCl₂ and balancing with equimolar substitution of NaCl.

Static incubation studies. To examine the long-term effects of agents on hormone release, cells were cultured in plating medium in 24-well dishes (Falcon 3847). Cells were incubated (28°C, 5% CO₂) for 24 h before drug testing. After a rinse in testing medium, cells were exposed to drugs in a static incubation setting (28°C, 5% CO₂) for 2 h (5). For 1,2-bis(2-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid-acetoxymethyl ester (BAPTA-AM) studies, cells were incubated for 40 min in BAPTA-AM at 28°C before addition of caffeine to allow time for concentration of the chelator intracellularly and cleavage of the ester bond. After a 2-h incubation, the testing medium was removed and stored at 20°C before the GH content was assayed by radioimmunoassay (39). Hormone release (ng/ml) over the 2-h incubation period was normalized to the mean basal value and compared using ANOVA. A Fisher's post hoc least-significant difference (PLSD) test was used for post hoc comparison. Significance was determined to be when P < 0.05. All data are expressed as means ± SE.

Column perifusion studies. For column perifusion studies, cells were cultured overnight on preswollen Cytodex-I (Sigma) beads, as described previously (5, 6). The next day, beads and cells were loaded into column perifusion chambers for drug testing. Testing medium was used for column perifusion at a flow rate of 15 ml/h and a temperature of 18°C. Cells were allowed a minimum 4-h wash before testing. Unless otherwise stated, perifusates were collected as 5-min fractions that were stored at 20°C until being sampled for GH content by specific radioimmunoassay (39).

Drugs. All drugs were made up in testing medium. Caffeine (Research Biochemicals International, Natick, MA) was dissolved in testing medium. Ryanodine and BAPTA-AM (Calbiochem, La Jolla, CA) were initially dissolved in DMSO as a 1,000-fold stock solution. Tetraethylammonium chloride (TEA-Cl), 4-aminopyridine (4-AP), and cadmium chloride (CdCl₂) were bought from Sigma Chemical (St. Louis, MO). TEA (5 mM) and 4-AP (2 mM) were made up immediately before use. CdCl₂ was prepared as a 50 mM stock solution in distilled deionized water and was diluted 1:1,000 to give a final concentration of 50 µM in testing solution. The PKA inhibitor (H-89) was purchased from Seikagaku and made up as stock aliquots of 10 mM in absolute ethanol. These aliquots were kept frozen and diluted at 1:1,000 immediately before use.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Caffeine is a potent GH secretagogue. Caffeine (10 mM) is a potent secretagogue for GH release (28). In static incubation, millimolar concentrations of caffeine elicited a dose-dependent increase in GH release (Fig. 1A). In column perifusion studies with a 1-min perifusate collection rate, 10 mM caffeine elicited a characteristic biphasic increase in hormone release with a peak value of ~800% (Fig. 1B) followed by a sustained/plateau phase.

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Caffeine stimulates growth hormone (GH) release from dispersed goldfish pituitary cells. A: dose-response curve of GH release in static incubation in response to increasing doses of caffeine (Caff; n = 6 wells/dose); expressed as %basal (0 Caff), which was 16.0 ± 3.3 ng/ml. Treatments that were not significantly different from each other [ANOVA followed by Fisher's post hoc least-significant difference (PLSD), P > 0.05] are indicated by an underscore. B: GH release in rapid perifusion (1-min fractions, n = 6 columns).

Is elevation of intracellular Ca²⁺ a necessary component for evoked secretion? Given the dramatic secretory responsiveness of GH cells to caffeine, we felt it necessary to determine the Ca²⁺ dependence of caffeine-evoked GH release. If caffeine stimulated GH release through a mechanism that elevated [Ca²⁺]_i, then blocking such a rise by using an intracellular chelator should block evoked release. Therefore, we preincubated cells for 40 min with 50 µM BAPTA-AM, which is a membrane-permeant form of the Ca²⁺ chelator BAPTA (K_D with Ca²⁺ = 2.33 nM). On entering into the cell, endogenous esterases cleave the ester bond concentrating the BAPTA inside the cell and making it available to chelate [Ca²⁺]_i. In static incubation, BAPTA-AM slightly suppressed but did not abolish basal secretion, suggesting that the cells had a largely intact constitutive release mechanism. BAPTA-AM blocked most of the caffeine-evoked release of GH (Fig. 2), supporting the fact that elevation of [Ca²⁺]_i is a necessary component for caffeine-evoked GH release.

View larger version (13K): [in this window] [in a new window]   Fig. 2.   Caff-evoked release is blocked by 1,2-bis(2-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid (BAPTA). GH release in static incubation in response to 10 mM Caff in the presence or absence of the membrane-permeant Ca2+ chelator BAPTA-acetoxymethyl ester (AM; 50 µM, 40-min preincubation); expressed as %basal (control), which was 13.7 ± 0.3 ng/ml (n = 15 or 16 wells for each treatment). Different symbols (*, **) indicate hormone-release values that are significantly different from each other. (ANOVA, followed by Fisher's PLSD, P < 0.05).

Does the ryanodine receptor mediate the caffeine response? Caffeine-evoked GH release is sensitive to TMB-8, suggesting that an intracellular Ca²⁺-release channel participates in the caffeine stimulation of hormone release (28). We therefore examined whether caffeine stimulates GH release through its actions on the ryanodine receptor. A two-pulse paradigm was used for two reasons. First, ryanodine stabilizes a subconductance open state of the channel, inhibiting the ability of the store to refill (2, 16, 17, 56) and therefore inhibiting subsequent Ca²⁺ release from stores. Second, ryanodine preferentially binds to the activated channel (54). Consequently, the ability of ryanodine to bind to the receptor and block the caffeine response would only be observed in the first pulses if the receptor was constitutively active.

View larger version (34K): [in this window] [in a new window]   Fig. 3.   Ryanodine (100 µM) reduces the peak response of Caff-evoked GH release. A: column perifusion of dispersed pituitary cells. Filled horizontal bar indicates time of application of 100 µM ryanodine. Empty horizontal bars indicate time of Caff (10 mM) application (basal hormone release was 13.2 ± 1.5 and 14.0 ± 0.7 ng/ml for Caff only or Caff + ryanodine, respectively, n = 4 columns for each treatment). B: quantified net responses for column perifusions depicted in A. *Significantly different responses between the first and second pulses within each treatment regime (paired t-test, P < 0.05).

Role of extracellular Ca²⁺ entry in caffeine-evoked hormone release. We further tested the sensitivity of caffeine-evoked release to Ca²⁺ by perturbing Ca²⁺ entry and determining the effects on the GH response to caffeine. First, we used Ca²⁺-free M-199 to determine if removal of extracellular Ca²⁺ would affect caffeine-elicited GH release (Fig. 4). Removal of extracellular Ca²⁺ caused a marked increase in GH secretion, and 15-min pretreatment with Ca²⁺-free solution completely blocked caffeine-evoked hormone release, suggesting that caffeine-evoked GH release is dependent on the availability of extracellular Ca²⁺.

View larger version (23K): [in this window] [in a new window]   Fig. 4.   Ca2+-free medium blocks Caff-evoked release. Column perifusion of Caff-evoked release in the presence or absence of Ca2+-free solution or regular M-199 (1.2 mM Ca2+). Note the slight increase in basal release in the presence of Ca2+-free medium. Basal release values were 24.7 ± 0.8, 23.8 ± 2.5, and 23.5 ± 3.8 ng/ml for Caff only (n = 4), Ca2+-free only (n = 3), and Caff in Ca2+ free, respectively (n = 4).

View larger version (24K): [in this window] [in a new window]   Fig. 5.   Caff-evoked hormone release is sensitive to blockade with Cd2+ in a time-dependent manner. A: column perifusion of Caff-evoked release with a 5-min pretreatment, no pretreatment of Cd2+, or no Cd2+ in the perifusate. Basal secretion values were 33.6 ± 1.8, 40.2 ± 4.8, and 35.8 ± 3.7 ng/ml for control, no pretreatment, and 5-min pretreatment columns, respectively (n = 6 columns/treatment). Downward arrow indicates end of Cd2+ treatment. B: quantified peak, plateau, and total net responses to different treatment times of Cd2+. Underscore indicates not significantly different among treatments (ANOVA, P > 0.05). Experiments (n = 6 columns/treatment) were performed in June. C: quantified peak, plateau, and total net responses to 5- and 35-min pretreatments with Cd2+. Experiments (n = 6 columns/treatment) were performed in March.

Dopamine elicits an increase in GH release whose initiation is dependent on entry of extracellular Ca²⁺ entry through high voltage-activated Ca²⁺ channels. The results described above suggested that caffeine-evoked GH release is initiated by mobilization of Ca²⁺ from an intracellular Ca²⁺ store that is dependent on HVCC activity for its maintenance. As it has been reported that DA-induced GH release is dependent on Ca²⁺ entry through HVCCs, it could be hypothesized that caffeine and DA stimulation of GH release might be related in their extracellular Ca²⁺ dependency. To evaluate this possibility, we examined the sensitivity of DA-induced GH release to different pretreatment regimes with the HVCC blocker Cd²⁺ as we had done in the previous caffeine studies (Fig. 5). Previous studies on HVCCs and extracellular Ca²⁺ involvement in DA D₁ action used prolonged (>20 min) preexposure to channel inhibitors (62, 63) and could not distinguish between an immediate involvement of HVCCs in activation of hormone secretion and the long-term requirement of Ca²⁺ entry to replenish intracellular Ca²⁺ stores. In the present study, applications of a 5-min pulse of 1 µM DA elicited a rapid increase in hormone release that was attenuated both by simultaneous application as well as a 5-min pretreatment with 50 µM Cd²⁺ (Fig. 6). Interestingly, whereas the initial (peak) GH-response phase was reduced by simultaneous Cd²⁺ treatment and further reduced with preexposure to Cd²⁺, the plateau was significantly reduced in the column that received no pretreatment with Cd²⁺.

View larger version (24K): [in this window] [in a new window]   Fig. 6.   Dopamine D1-evoked GH release is immediately blocked by Cd2+ (50 µM). A: column perifusion study of GH release on 5 min of application of dopamine (1 µM) in the presence of Cd2+ with 5 min or no pretreatment. Basal secretion values were 37.5 ± 3.7, 32.5 ± 1.4, and 43.7 ± 1.5 ng/ml for dopamine, no pretreatment, and 5-min pretreatment columns, respectively. Downward arrow indicates removal of Cd2+. B: quantified net responses for peak, plateau, and total net GH release. Underscore denotes not significantly different values (P > 0.05, n = 6 columns/treatment).

Interaction between cyclic nucleotides and caffeine-evoked release. One putative action of caffeine is to inhibit phosphodiesterases resulting in an accumulation of cyclic nucleotides. To further clarify the mechanism of action of caffeine, cells were pretreated with a potent and selective inhibitor of protein kinase A, H-89 (10 µM). Preexposure to H-89 for 35 min reduced GH release to 55.2 ± 7.6% of the initial basal release (Fig. 7A, control column 99.5 ± 7.1%). Moreover, the quantified net GH response to caffeine was reduced to only 31% of control values (Fig. 7, A and B). Notable is the fact that a major portion of the total reduction by H-89 is due to the significant decrease in the peak response (unpaired t-test, P < 0.05), whereas the prolonged plateau phase is not significantly affected. These results suggest that there is a high degree of interaction between caffeine's effects and the cAMP/PKA signaling pathway in mediating the peak secretory response in somatotropes.

View larger version (23K): [in this window] [in a new window]   Fig. 7.   The protein kinase-A (PKA) inhibitor H-89 blocks Caff-induced GH release. A: cells were exposed to 35 min of pretreatment of 10 µM H-89 before application of a 25-min pulse of 10 mM Caff. Basal values of GH release were 35.1 ± 3.0, 30.9 ± 3.9, and 28.2 ± 1.1 ng/ml for each 5-min fraction for columns exposed to Caff only, H-89 only, or both, respectively. Experiments were performed in September during times of gonadal regression. B: quantified net responses Caff-evoked GH release. *Significance between hormone-release responses in the presence or absence of H-89 (n = 6 columns/treatment, unpaired t-test, P < 0.05).

Other nonspecific actions of caffeine. It has been shown previously that caffeine may inhibit K⁺ channels by either direct or indirect mechanisms (1, 47, 49). Such a blockade could also explain caffeine-evoked release, because the decrease of a voltage-dependent K⁺ conductance would have the effect of depolarizing the cells, increasing Ca²⁺ entry through HVCCs, and thereby increasing [Ca²⁺]_i and hormone release. To examine this possibility, we used the general K⁺-channel blocker TEA (5 mM) and the more selective A-current blocker 4-AP (2 mM) in column perifusion studies to determine whether blockade of K⁺ channels was involved in caffeine's mechanism of action. These concentrations were selected based on their ability to block K⁺ conductances in goldfish pituitary cells (45, 60). Despite 35 min of pretreatment, TEA had no effect on the net GH response to caffeine (control: 328 ± 42% vs. TEA: 326 ± 40%, P > 0.05). Similarly, 4-AP also had no effect (control: 725 ± 51% vs. 4-AP: 833 ± 79%, P > 0.05), suggesting that any putative caffeine-mediated inhibition of K⁺ channels is not involved in the GH secretory response to caffeine in goldfish somatotropes.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The present study characterized caffeine's actions on hormone release in goldfish somatotropes and elucidated some of the properties of caffeine-sensitive mechanisms in stimulation of GH secretion. Current results extend previous work that showed that goldfish somatotropes have a potent secretory mechanism that is readily stimulated by caffeine (28). We demonstrate that caffeine-evoked GH release from somatotropes is dependent on elevation of [Ca²⁺]_i as it can be blocked by pretreatment of the membrane-permeant Ca²⁺ chelator BAPTA-AM. This powerful secretory response that has been shown to be sensitive to TMB-8 (28) is also partially ryanodine sensitive, being reduced in a use-dependent manner. Moreover, we show that caffeine, in contrast to DA, stimulates GH secretion using a signal-transduction cascade that only indirectly involves HVCCs.

Goldfish somatotropes possess at least two caffeine-activated intracellular Ca²⁺ stores that vary in their dependency on HVCCs for maintenance of filling state. We propose that there is an initial trigger store that is highly sensitive to blockade of Ca²⁺ channels. On the basis of the fact that the secretory response is reduced by 5 min of pretreatment with Cd²⁺, we hypothesize that Ca²⁺ release from this store is relatively leaky and is depleted rapidly (<5 min). Such a rapidly depleting store seems similar in this regard to the IP₃-activated store in rat gonadotropes that, when extracellular Ca²⁺ is removed, is depleted within 5-10 min (58). Therefore, the present study suggests that Ca²⁺ entry through HVCCs is an important mechanism of maintenance for the amount of Ca²⁺ available for release from this caffeine-sensitive "trigger" store in goldfish somatotropes. This store mediates >50% of the peak phase of the GH response to caffeine. In addition, somatotropes possess stores that are not inhibited by prolonged blockade of HVCCs and that mediate a portion both of the peak and the plateau secretory responses to caffeine.

As ryanodine sensitivity is restricted to the peak phase of the GH response to caffeine, the ryanodine receptor may mediate a portion of release from this trigger store. The ryanodine sensitivity of caffeine's actions is well documented in a variety of cell types (43), including rat somatotropes (21, 44). The ability of ryanodine to block caffeine action is use-dependent as it binds preferentially to the open channel (54) and may stabilize the channel in a subconductance state (48), impairing the ability of stores to refill after discharge (2, 16, 17, 56). Here, we show that sequential application of caffeine in the presence of ryanodine causes a reduced GH response, implicating the presence of a use-dependent block of the caffeine store. Whether this ryanodine-sensitive, caffeine-activated intracellular Ca²⁺ store is targeted by other endogenous activators associated with caffeine/ryanodine, such as cyclic-ADP ribose (19, 30) and acyl-coenzyme A (14), will need to be examined further.

Application of H-89 also substantially reduced the peak phase of the GH response to caffeine, suggesting that inhibition of a constitutively active phosphodiesterase might be involved in a portion of the GH response. Consistent with this idea, 3-isobutyl-1-methylxanthine is a potent stimulator of GH release (61). Because the peak portion is also partially sensitive to ryanodine, PKA and ryanodine may be closely associated in the initiation of GH release in response to caffeine. Indeed, recent studies in pancreatic B cells have shown that PKA exerts important facilitatory actions on the ryanodine receptor and, in fact, may be necessary for its activation (25, 26). Preliminary results in goldfish somatotropes suggest that forskolin-evoked release is also reduced by ryanodine, further reinforcing the interaction between cAMP/PKA signaling and the ryanodine receptor (J. D. Johnson and J. P. Chang, unpublished observations). However, H-89 pretreatment may also reduce any resting cAMP/PKA activity, thereby affecting HVCC activity and subsequent Ca²⁺ store-filling state.

Indeed, as has been shown previously in static incubation (62, 63), the current perifusion studies demonstrate that application of H-89 or blockade of HVCCs reduces basal secretion, suggesting that HVCCs and regulation of PKA activity are critical in the control of basal secretion in perifused goldfish somatotropes. Preliminary imaging studies suggest that Cd²⁺ treatment does not reduce basal [Ca²⁺]_i in identified somatotropes (C. J. Wong and J. P. Chang, unpublished observations). Therefore, a global reduction in [Ca²⁺]_i is likely not the mechanism by which Cd²⁺ exerts its inhibitory action on basal or evoked release. However, Cd²⁺ blockade may block local Ca²⁺ transients arising from spontaneous openings of any HVCCs that may be in close apposition to vesicles and that might mediate constitutive release.

The indirect involvement of HVCCs in caffeine-evoked release complements the fact that intracellular Ca²⁺ stores are the primary mechanism of action mediating caffeine-evoked secretion. Approximately 28% of the peak component of caffeine-stimulated GH release is blocked by 100 µM ryanodine, whereas up to 80% of the total response is blocked by TMB-8 (28). Therefore, there appears to be an additional class or classes of ryanodine-insensitive intracellular channels that mediate the bulk of the GH secretory response to caffeine. Such novel release channels that are activated by caffeine but are ryanodine-insensitive have been identified in pancreatic acinar cells (10, 50) and may be part of a complex intracellular network of interacting stores (28). However, the location and morphological correlates of these novel caffeine-sensitive sites in goldfish somatotropes are, at present, unknown.

Although both extremely potent secretagogues and although both involve Ca²⁺ entry and cAMP/PKA, DA and caffeine actions have very different dependencies on Ca²⁺ entry through HVCCs. Whereas previous studies have shown that D₁-evoked GH release was impaired by prolonged pretreatment with Ca²⁺-channel blockers (62, 63), it is clear from this study that simultaneous application of Cd²⁺ and DA results in a decreased secretory response. Unlike caffeine, which requires HVCCs only to maintain a subset of its stores for its secretory response, HVCC activation and its consequences are critical parts of the early signaling events in acute DA-evoked GH release. Additional lines of evidence support the hypothesis that cAMP/PKA activation of HVCCs mediates the GH secretory response to DA in goldfish somatotropes: DA stimulation causes an increase in cAMP release in goldfish pituitary cells (61); DA enhances inward Ca²⁺ currents in identified goldfish somatotropes (7); cAMP-dependent phosphorylation is well known to increase the probability of channel reopening for both native and expressed dihydropyridine-sensitive channels (27, 29) resulting in enhanced whole cell currents (13). Therefore, it is likely that the initial Cd²⁺-sensitive phase of DA's actions is mediated by a D₁-stimulated, cAMP/PKA-mediated increase in an L-type Ca²⁺ conductance, similar to what has been found in chromaffin cells (3, 4). Although it is unknown whether D₁ activation of Ca²⁺ channels is linked to subsequent CICR in goldfish somatotropes, the GH response to DA is reduced by TMB-8 (J. D. Johnson and J. P. Chang, unpublished observations), implicating an intracellular Ca²⁺-release channel in a portion of DA-stimulated GH secretion. Moreover, DA-evoked GH release may have several additional components of activation that are distal to cAMP/PKA and independent of any putative HVCC activation such as intracellular channel activation (see above) or vesicle packaging, docking, and priming (22).

Despite of the proposed differential involvement of HVCCs in mediating the initial secretory responses to caffeine compared with DA, the GH responses both to DA and caffeine had significant components that were insensitive to blockade of HVCCs by Cd²⁺. It is unclear as to whether both DA and caffeine share similar Cd²⁺-insensitive secretory mechanisms. Moreover, the origin of the Cd²⁺-insensitive component is unknown. Goldfish somatotropes possess only high voltage-activated Ca²⁺ channels (8) and do not possess a low-threshold voltage-activated transient or T current, unlike their mammalian homologs (11, 42). Therefore, any response mediated by a Cd²⁺-insensitive voltage-gated Ca²⁺ channel can be excluded. However, the secretory response to caffeine was abolished by Ca²⁺-free medium. There are three possible hypotheses for this effect. The first is that the residual Cd²⁺-insensitive secretory response to caffeine is mediated by a store-operated channel that is coupled to exocytosis similar to what has been identified in chromaffin cells (15). Therefore, removal of extracellular Ca²⁺ would abolish Ca²⁺ entry through these channels. A second hypothesis is that removal of extracellular Ca²⁺ may prevent maintenance of an intrinsically leaky caffeine-activated store, thereby limiting the ability of caffeine to evoke a response. The last alternative is that removal of extracellular Ca²⁺ may induce the discharge of a caffeine-sensitive store, thereby blocking the ability of caffeine to further release Ca²⁺ from this store.

In support of the latter hypothesis is the fact that GH secretion increases on application of Ca²⁺-free medium (28, and this study), differing strongly from selective blockade of Ca²⁺ channels with Cd²⁺ or PKA inhibition, both of which reduce GH secretion (61-63, and this study). Although this requires further examination with Ca²⁺ imaging, it is possible that goldfish somatotropes possess a novel Ca²⁺-homeostatic mechanism that discharges intracellular stores in the absence of extracellular Ca²⁺.

In conclusion, previously, it has been shown that GnRH and DA stimulation of GH release in goldfish both require HVCCs despite having other divergent signal-transduction components (63). Here, we demonstrate that caffeine-activated stores, which partially overlap with those targeted by GnRH, and DA signaling have a differential sensitivity to blockade of HVCCs. In particular, DA has a proximal requirement for Ca²⁺ channels in the initiation of secretion, whereas a subset of the caffeine/GnRH stores indirectly rely on them for maintenance and subsequent generation of a secretory response. Finally, we provide evidence for the presence of multiple intracellular stores that are activated by caffeine, yet that exhibit different sensitivities to blockade of HVCCs ranging from rapidly depleteing (<5 min) to insensitive to HVCC blockade.

Perspectives