The Expression and Role of Hyperpolarization-Activated and Cyclic Nucleotide-Gated Channels in Endocrine Anterior Pituitary Cells

Abstract

Pituitary cells fire action potentials independently of external stimuli, and such spontaneous electrical activity is modulated by a large variety of hypothalamic and intrapituitary agonists. Here, we focused on the potential role of hyperpolarization-activated and cyclic nucleotide-gated (HCN) channels in electrical activity of cultured rat anterior pituitary cells. Quantitative RT-PCR analysis showed higher level of expression of mRNA transcripts for HCN2 and HCN3 subunits and lower expression of HCN1 and HCN4 subunits in these cells. Western immunoblot analysis of lysates from normal and GH₃ immortalized pituitary cells showed bands with appropriate molecular weights for HCN2, HCN3, and HCN4. Electrophysiological experiments showed the presence of a slowly developing hyperpolarization-activated inward current, which was blocked by Cs⁺ and ZD7288, in gonadotrophs, thyrotrophs, somatotrophs, and a fraction of lactotrophs, as well as in other unidentified pituitary cell types. Stimulation of adenylyl cyclase and addition of 8-Br-cAMP enhanced this current and depolarized the cell membrane, whereas 8-Br-cGMP did not alter the current and hyperpolarized the cell membrane. Both inhibition of basal adenylyl cyclase activity and stimulation of phospholipase C signaling pathway inhibited this current. Inhibition of HCN channels affected the frequency of firing but did not abolish spontaneous electrical activity. These experiments indicate that cAMP and cGMP have opposite effects on the excitability of endocrine pituitary cells, that basal cAMP production in cultured cells is sufficient to integrate the majority of HCN channels in electrical activity, and that depletion of phosphatidylinositol 4,5-bisphosphate caused by activation of phospholipase C silences them.

Pituitary cells fire action potentials (AP) independently of external stimuli, a phenomenon termed “spontaneous electrical activity.” Each AP is composed of a slow depolarizing phase, a rapid depolarizing phase or spiking depolarization, and a rapid or delayed (plateau-bursting type) repolarizing phase (1). Such rhythmicity fulfills the need to drive the periodic fluctuations in cytosolic Ca²⁺ concentrations and hormone release, as is well documented for lactotrophs and somatotrophs (2). In general, intrinsic electrophysiological characteristics of neuronal, neuroendocrine, and muscle cells reflect the type and density of numerous voltage- and ligand-gated ion channels that regulate the flow of ionic currents across the plasma membrane (3). In that respect, it is not unexpected that endocrine pituitary cells also possess a rich repertoire of ion channels, including voltage-gated Na⁺, K⁺, and Ca²⁺ channels, ligand-gated GABAγ-aminobutyric acid-A, and purinergic P2X receptor channels, as well as a number of less conventional ionic conductances. Spontaneous electrical activity of these cells is modulated by a large variety of hypothalamic and intrapituitary agonists and intracellular messenger systems. These include agonists for the G protein-coupled receptors; the G_s-coupled receptors (CRH, GHRH, vasoactive intestinal peptide/pituitary adenylate cyclase-activating polypeptide) facilitate firing of AP, the G_i/o-coupled receptors (dopamine and somatostatin) inhibit it, and the G_q/11-coupled receptors (GnRH, TRH, arginine vasopressin) have dual roles, an initial inhibitory and a sustained stimulatory role (1).

In neurons and cardiac cells, the ionic channels that generate autonomous pacemaking capabilities are frequently members of the small family of pacemaking channels. These channels, termed “hyperpolarization-activated and cyclic nucleotide-gated” (HCN) channels, belong to the superfamily of voltage-gated channels but form a distinct subgroup of channels that are closely related to voltage-independent cyclic nucleotide-gated (CNG) channels (4). Unique voltage dependence of HCN channels, together with permeability for both K⁺ and Na⁺, explains their pacemaking function in cardiac tissue (5, 6). HCN and CNG channels also play important roles in spontaneous and receptor-induced excitability of other cells by increasing the slope of slow depolarization. This comes from their regulatory properties. The hyperpolarization-activated cation current, termed “I_h” (h stands for hyperpolarization), is sensitive to the presence of cAMP and, to a much weaker extent, cGMP. Cyclic nucleotides not only accelerate the kinetics of activation of I_h, but also shift the voltage dependence for activation toward more depolarized values. On the other hand, CNG channels expressed in photoreceptors have a strong preference for cGMP, whereas the olfactory channel is almost equally sensitive to both ligands (4). The dependence of HCN and CNG channel activation on cyclic nucleotides provides a rationale for the stimulatory effects of G_s-coupled receptors and nitric oxide-soluble guanylyl cyclase signaling pathways on the electrical activity in excitable cells. The G_i/o-dependent inhibitory actions on cAMP-mediated regulation of I_h have also been documented (7). The G_q/11-coupled receptors also influence the gating of HCN and CNG channels by affecting the phosphatidylinositol 4,5-bisphosphate (PIP₂) levels (8).

To date, four mammalian channel subunits, termed “HCN1–4,” have been cloned. These subunits, organized as homo- and heterotetrameric complexes, represent the molecular correlates of the I_h (9, 10). The pharmacological identification of these channels is based on the sensitivity to several inhibitors, including ZD7288 and extracellular Cs⁺, and insensitivity to extracellular Ba²⁺ (11, 12). In vertebrates, there are six CNG subunits: CNGA1, CNGA2, CNGA3, CNGA4, CNGB1, and CNGB3. CNGA1–3 subunits can form homomeric channels in heterologous expression systems, and the other subunits can coassemble to form functional heteromeric channels (4). The mRNA transcripts and functional HCN channels were identified in rat GH₃ cells and somatotrophs (13–15), mouse At-T20 cells (16), and frog melanotrophs (17), but the expression and role of HCN channels in electrical activity in other endocrine pituitary cells were not studied. The mRNA transcripts for CNG channels were also identified in rat pituitary cells (18), whereas the dependence of pituitary cell excitability on cAMP vs. cGMP was not clarified. The influence of the G_q/11 signaling pathway on the activity of these channels was also not studied.

Here we focus on the expression and role of HCN channels in electrical activity in cultured somatotrophs, gonadotrophs, lactotrophs, and thyrotrophs. The biophysical properties of I_h were systematically investigated in these cell types, including the slow kinetics and voltage dependence of their activation and the lack of inactivation. We also examined the pharmacological properties of these channels. Because there is not a common G_s-coupled receptor among these cells, we used forskolin, an activator of adenylyl cyclase, and cell-permeable 8-Br-cAMP and 8-Br-cGMP to activate these channels. Similarly, we down-regulated basal cAMP production and cAMP-dependent channel activity by inhibiting adenylyl cyclase. We also used pharmacological manipulations to alter the status of phospholipase C activity and levels of phosphoinositides. Single cell recordings were used in electrophysiological studies, and cyclic nucleotides were measured in mixed pituitary cell populations.

Results

Effects of cyclic nucleotides on electrical activity in anterior pituitary cells

In secretory anterior pituitary cells, both forskolin and cAMP stimulate electrical activity. Figure 1, A and E, shows stimulatory effects of cell permeable 8-Br-cAMP on electrical activity in thyrotrophs and gonadotrophs, respectively, and Fig. 1D shows effect of forskolin on electrical activity in gonadotrophs. In contrast to the depolarizing effect of 8-Br-cAMP, 8-Br-cGMP hyperpolarized the plasma membrane in thyrotrophs (Fig. 1, B and C), which slowed the firing frequency in spontaneously firing cells (Fig. 1B). Similar effect of 8-Br-cGMP was also observed in other pituitary cell types. In addition, an 85-fold increase in cGMP production was caused by a 30-min incubation with 10 μm 3,3′-(hydroxynitrosohydrazino)bis-1-propanamine, a nitric oxide donor (basal cGMP was at 44 ± 8 fmol/well, n = 6, 3,3′-(hydroxynitrosohydrazino)bis-1-propanamine-stimulated cGMP was at 3.77 ± 0.1 pmol/well, n = 6). However, this incubation was still unable to stimulate electrical activity in anterior pituitary cells (data not shown). These results indicate that HCN channels, but not CNG channels, could account for cAMP-stimulated electrical activity in endocrine pituitary cells, whereas cGMP inhibits electrical activity through other channels.

Fig. 1.

Expression of HCN cation channels in rat pituitary cells. A–C, Effects of 8-Br-cAMP (A) and 8-Br-cGMP (B and C) on electrical activity in thyrotrophs. Note both the depolarizing effect of cAMP and the hyperpolarizing effect of 8-Br-cGMP. D and E, Forskolin-(D) and 8-Br-cAMP-stimulated (E) stimulated electrical activity in pituitary gonadotrophs. F, Quantitative RT-PCR analysis of HCN subunit mRNA transcript expression in anterior pituitary cells cultured for 24 h. G, Western blot analysis of HCN expression in primary and GH₃ immortalized pituitary cells. s, Seconds.

Expression of HCN channels

To clarify the expression of HCN channels in endocrine pituitary cells, we performed quantitative RT-PCR analysis. This analysis confirmed the presence of mRNA transcripts for HCN1, HCN2, HCN3, and HCN4 in mixed populations of rat anterior pituitary cells cultured for 24 h. The real-time quantitative PCR and the comparative Ct method (19) also revealed that HCN2 mRNA transcripts are the most abundantly expressed, followed by HCN3, HCN4, and HCN1 (Fig. 1F). The presence of HCN isoforms was also studied in mixed rat pituitary cells and GH₃ immortalized lactosomatotrophs using Western blot analysis. We were unable to detect any specific band using HCN1 antibodies. In contrast, double immunoreactive bands for HCN2 and HCN3 were detected, consistent with the presence of glycosylated and nonglycosylated forms, whereas HCN4 subunit was observed as a single immunoreactive band (Fig. 1G). Moreover, all immunopositive bands appeared at different molecular weights, and their positions were in good agreement with earlier data obtained from other tissues expressing HCN channels (20, 21). Finally, rat liver tissue was used as negative control as no mRNA expression of HCN isoforms was detected (results not shown), in accordance with earlier published data (20, 21).

Biophysical characterization of native HCN channels

Experiments were performed in cultured single pituitary cells 24 h after dispersion. Somatotrophs were identified by their responses to GHRH. As recently described, these cells respond with variable patterns of electrical activity (22). Figure 2 shows that GHRH induced single spiking in a quiescent somatotroph (panel A) and increased the frequency of plateau-bursting type of electrical activity in a spontaneously active cell (panel B). Both lactotrophs and thyrotrophs express TRH receptors, but in contrast to lactotrophs, thyrotrophs do not express dopamine D2 receptors. Figures 3A and 4A, show the pattern of electrical activity triggered by TRH in lactotrophs and thyrotrophs, respectively; the former also responded to the application of bromocriptine, a dopamine D₂ agonist, by hyperpolarization of the plasma membrane (Fig. 3A, bottom). Note that TRH triggered electrical activity with the peak AP amplitude reaching +30 to +40 mV in thyrotrophs (Fig. 4A), whereas AP overshooting was not observed in lactotrophs (Fig. 3A). Gonadotrophs were identified by their oscillatory and apamin-sensitive response to 1 nm GnRH (23, 24) (Fig. 5A). Cells not responding to these agonists were described as unidentified.

Fig. 2.

Characterization of I_h current in pituitary somatotrophs. A and B, Identification of somatotrophs by GHRH-induced electrical activity in a quiescent cell (A) and modulation of the frequency of AP in a spontaneously active cell (B). C, Representative example of the whole-cell current response to a hyperpolarizing voltage step to −120 mV from a holding potential of −40 mV. D, In all somatotrophs (S), the channel opening was best described by a single-exponential fit. In gonadotrophs (G), lactotrophs (L), and thyrotrophs (T), both single- and double-exponential developments of current were observed. E, Activation curve for I_h. Tail current measurements were used. Cells were held at a holding potential of −40 mV and pulsed in 20-mV increments to test potentials between −60 and −120 mV for 7.5 sec. Normalized amplitudes of tail currents I/I_max were plotted against testing voltage and fitted with the Boltzmann equation (see “Calculations”). Averaged data (means ± sem) from seven cells are shown. F and G, Representative traces of I_h current before and 10 min after addition of 100 μm ZD7288 (F) and before and after application of 1 mm Cs⁺ (G) for 30 sec. H, Reduction in AP frequency by 1 mm Cs⁺ in a spontaneously firing cell expressing I_h current. s, Seconds.

Fig. 3.

Characterization of I_h current in rat pituitary lactotrophs. A, Lactotrophs were identified by their sensitivity to TRH (top) and bromocriptine (bottom). B and D, Whole-cell voltage-clamp recordings of I_h in identified lactotrophs in the presence (gray) and absence (black) of 1 mm Cs⁺ (B) and ZD7288 (D). C, The majority of lactotrophs do not express I_h (top) or show small amplitude current response (middle). Only about 30% of lactotrophs express I_h with amplitude comparable to other anterior pituitary cells (bottom). E, The activation curve for I_h obtained by tail current analysis. F (main panel), Dose-dependent effects of MDL12330A, an adenylyl cyclase inhibitor, on cAMP release in a mixed population of pituitary cells. Inset, Bath application of MDL12330A reduced the amplitude of I_h in an identified lactotroph.

Fig. 4.

Characterization of I_h current in rat pituitary thyrotrophs. Panel A, Thyrotrophs were identified by their response to TRH and their lack of response to bromocriptine. Note the difference in the peak amplitude of AP in thyrotrophs and lactotrophs (Figs. 3A vs. 4A). Panel B, Representative example of the whole-cell current response to a hyperpolarizing voltage step to −120 mV from a holding potential of −40 mV. The time course of channel opening was best described by a double-exponential fit. Panel C, Inhibition of I_h current by 10 μm ZD7288. D, Stimulation of I_h by 8-Br-cAMP. E, Inhibition of I_h by MDL12230A. Top, Representative trace; bottom, Normalized current- and time-constant (τ) values (mean ± sem).s, Seconds; C, control.

Fig. 5.

Properties of HCN channels in pituitary gonadotrophs. A, Cells were identified by their oscillatory response to 1 nm GnRH applied at the end of the recording. B, Representative example of the whole-cell current response to a hyperpolarizing voltage step to −120 mV from a holding potential of −40 mV. The time course of channel opening was best described by a single-exponential fit. C, The activation curve for I_h was obtained by tail current analysis as described in Fig. 2E. D, Blockade of I_h current by the addition of 100 μm ZD7288 to extracellular solution. ZD7288 was applied to the bath for 10 min to fully develop its blocking effect. E, Inhibition of I_h by 1 mm bath Cs⁺. F and G, Cells expressing I_h (F, bottom) displayed inward rectifications in response to hyperpolarizing current pulses of −5 pA that were suppressed by 1 mm Cs⁺ (F, top) and were absent (G, top) in cells lacking I_h (G, bottom). H and I, Difference in the effects of bath Cs⁺ on the frequency of spontaneous firing of AP in gonadotrophs expressing (H) and not expressing (I) I_h. Percentage decrease (H, downward arrow) or increase (I, upward arrow) in frequency is indicated for each cell group. s, Seconds.

In these cell types, the hyperpolarization of the plasma membrane resulted in activation of a slowly developing inward current in endocrine pituitary cells held at −40 mV and bathed in medium containing 20 mm K⁺. The current started to appear at voltages approaching −60 mV, and both the amplitude and the rate of activation of this current increased with further hyperpolarization. Representative examples of the whole-cell current response to a hyperpolarizing voltage step to −120 mV from a holding potential of −40 mV are shown in Fig. 2C (somatotrophs), Fig. 3C (lactotrophs), Fig. 4B (thyrotrophs), and Fig. 5B (gonadotrophs). Such currents were observed in the majority of somatotrophs, thyrotrophs, and gonadotrophs (Table 1), as well as in unidentified pituitary cells (data not shown). In contrast, the majority of lactotrophs do not express I_h (Fig. 3C, top) or show small amplitude current response (middle), whereas only about 30% of lactotrophs expresses currents with amplitudes comparable to that recorded from somatotrophs and gonadotrophs (Fig. 3C, bottom; Table 1).

Table 1.

Biophysical properties of I_h in anterior pituitary cells

	Gonadotrophs	Lactotrophs	Somatotrophs	Thyrotrophs
Cells with Ih (%)	73.8	33.3	92.3	81.3
Amplitude (pA)	6.4 ± 0.7 (45)	8.5 ± 2.2 (31)	6.6 ± 1.1 (36)	15.9 ± 3.4 (8)
Amplitude (pA/pF)	1.1 ± 0.1 (43)	1.5 ± 0.4 (31)	1.7 ± 0.3 (35)	2.8 ± 0.6 (8)
τ (s)	0.9 ± 0.2 (38)	1.2 ± 0.2 (25)	1.7 ± 0.1 (35)	1.2 ± 0.2 (7)

The I_h was augmented by elevation of K⁺ in extracellular solution to 20 mm before recording. The I_h current was activated by voltage steps from −40 to −120 mV. Data shown are mean ± sem values and number in parentheses indicate number of cells.

The increase in current in somatotrophs was always fit best with a single-exponential function, whereas in other cell types the best approximation for the increase in current was achieved using single- or double-exponential fits with a time constant of activation ranging between 0.15 and 5 sec (Fig. 2D). The development of current was fastest in gonadotrophs and slowest in somatotrophs. The peak amplitude of current was comparable in gonadotrophs, lactotrophs, and somatotrophs and was doubled in thyrotrophs (Table 1). The currents did not inactivate significantly during the sustained application of a hyperpolarizing pulse (5–7 sec) but did deactivate rapidly after the end of the pulse.

Using the tail current measurements, we also obtained the activation curve of the I_h current in somatotrophs (Fig. 2E), lactotrophs (Fig. 3E), and gonadotrophs (Fig. 5C). In somatotrophs, a mean V_1/2 value of −93.2 ± 3.4 mV and a slope value of 9.3 ± 1.1 mV (n = 7), in lactotrophs a V_1/2 value of −94.1 ± 2.7 mV and a slope value of 8.6±1.2 mV (n = 4), and in gonadotrophs a V_1/2 value of −83.2 ± 3.3 mV and a slope value of 8.3 ± 1.2 mV were recorded. Such kinetic properties of the hyperpolarization-activated inward currents in anterior pituitary cells were similar to the properties of I_h currents present in neuronal and cardiac cells (11) as well as in GH₃ pituitary cells (13, 14).

Pharmacological characterization of native HCN channels

In all cell types, the current development was blocked by 100 μm ZD7288, an organic blocker of I_h (25, 26), and by 1 mm Cs⁺-containing medium. Consistent with the literature, the inhibitory effect of Cs⁺ in our experiments was instantaneous and reversible, whereas ZD7288 had to be applied for 10 min to achieve full and irreversible inhibition of this current. Effects of ZD7288 on the current amplitude are shown in somatotrophs (Fig. 2F), lactotrophs (Fig. 3D), thyrotrophs (Fig. 4C), and gonadotrophs (Fig. 5D). The effect of ZD7288 on the current did not differ between cell types. In gonadotrophs, ZD7288 inhibited 86 ± 6% (n = 4) of the current when applied at 100 μm concentration for 10 min, 62 ± 5% at 10 μm concentration for 10 min (n = 7), and 34 ± 4% when applied at 10 μm concentration for 5 min (n = 6). Effects of 1 mm Cs⁺ on the I_h amplitude are shown in Figs. 2G, 3B, and 5E. In all cell types, Cs⁺ (1 mm) inhibited 98 ± 1% of current when applied for 30 sec (n = 16). Gonadotrophs expressing I_h (Fig. 5F, bottom) displayed inward rectifications in response to hyperpolarizing current pulses of −5 pA that were suppressed by 1 mm Cs⁺ (5F, top) and were absent (5G, top) in cells lacking I_h (5G, bottom). Similar effects of Cs⁺ were observed in the three other cell types (data not shown).

Because of the nonspecific stimulatory effect of ZD7288 on electrical activity in pituitary cells described earlier (27), we could not use this compound in evaluating the role of HCN channels in spontaneous electrical activity. Application of 1 mm Cs⁺ not only inhibits HCN channels effectively, but also silences spontaneously active inward rectifier K⁺ channels in many cell types, including pituitary cells (28, 29). Consistent with the second effect, we observed an increase in the firing frequency of 26 ± 13% in cells not expressing I_h (n = 13). Figure 5I illustrates the stimulatory effect of 1 mm Cs⁺ on the firing frequency in a gonadotroph (n = 17) in which no I_h current was detected (5I, bottom). Figures 2H and 5H illustrate inhibitory effects of Cs⁺ on the firing frequency of I_h-expressing somatotrophs and gonadotrophs, respectively. Nonetheless, 1 mm Cs⁺ did not abolish spontaneous firing of AP in any of the cells expressing I_h. The effect of Cs⁺-induced inhibition of HCN channels on the frequency of AP is probably more pronounced but may be masked by its concomitant inhibition of classical inward rectifier K⁺ channels, which are also expressed in pituitary cells (28, 29). Notwithstanding, these results suggest that HCN channels contribute to, but are not critical for, spontaneous electrical activity in cultured pituitary cells.

Regulation of HCN channels

The ability of Cs⁺ to decrease the firing frequency suggested that basal cAMP levels in cultured pituitary cells are sufficient to activate these channels. To clarify this hypothesis, we treated pituitary cells with MDL12330A (cis-N-(2-phenylcyclopenthyl)azacyclotridec-1-en-2amine HCl), an inhibitor of the adenylyl cyclase family of enzymes. Figure 3F, main panel, illustrates the concentration-dependent effect of MDL12330A on basal cAMP production in cultured pituitary cells. MDL12330A (75 μm) also had a pronounced effect on I_h in lactotrophs (Fig. 3F, inset), thyrotrophs (Fig. 4E, top) and other pituitary cell types (data not shown). Figure 4E (bottom) illustrates that the current amplitude was reduced by 53 ± 8% (P < 0.0005, n = 10), and the current development was delayed for 222 ± 30% (P < 0.005, n = 9) in thyrotrophs treated with MDL12330A. The activation curve for I_h current was also measured before and after MDL12330A application. MDL12330A shifted the V_1/2 of activation curve to the left (before MDL12330A: −92.5 ± 0.8 mV; after MDL12330A: −105.8 ± 5 .9 mV) without change in slope factor (s) (s = 7.8 ± 1.8 mV, before MDL12330A; s = 7.7 ± 1.0 mV, after MDL12330A).

Because there is not a common G_s-coupled receptor among these cells, we studied effects of forskolin, an activator of adenylyl cyclase, and 8-Br-cAMP, a cell-permeable cAMP analog, on the peak amplitude of I_h and the rate of current development. I_h was measured before (control) and approximately 5 min after application of 1 μm forskolin or 1 mm 8-Br-cAMP. In contrast to MDL12330A treatment, forskolin slightly but significantly increased the current amplitude (111 ± 3% of control, P < 0.01, n = 7) and facilitated development of current (τ = 66 ± 8% of control, P < 0.005; n = 7). 8-Br-cAMP also increased the amplitude of I_h to 112 ± 6% of control and decreased the τ to 66 ± 9% of control (n = 7). In particular, the effect of 8-Br-cAMP on the amplitude of I_h was rather minor in somatotrophs (3.4% increase) and much greater in gonadotrophs (14.8%) and thyrotrophs (36.8%; Fig. 4D), whereas 8-Br-cAMP-induced decrease of τ was comparable among these different cell types. These results indicate that basal cAMP production in pituitary cells in vitro is sufficient to activate HCN channels and that forskolin can further facilitate gating of these channels. Difference in the sensitivity of I_h to modulation by cAMP could reflect the cell type-specific expression of different HCN subunits and/or variations in basal cAMP production among subpopulations of pituitary cells.

In gonadotrophs, we consistently observed attenuation of I_h current after application of GnRH (1 nm; Fig. 6D) that could be only slightly recovered by application of forskolin or 8-Br-cAMP (data not shown). On average (n = 16), GnRH decreased the peak amplitude of I_h to 76 ± 6% of control (P < 0.0005) and increased τ to 228 ± 39% (P < 0.005). The GnRH receptor belongs to the G_q/11 family of G proteins and is coupled to phospholipase C. Because it has been shown that HCN channels are allosterically modulated by phosphoinositides, we tested the sensitivity of pituitary HCN channels to phosphoinositides by inhibiting the activity of phosphoinositide 3-kinase, the enzyme responsible for their synthesis. Treatment of pituitary cells with the phosphoinositide 3-kinase inhibitor wortmannin (10 μm for >10 min) decreased the I_h current amplitude to 83 ± 5% of controls (P < 0.05) and increased the τ value to 190 ± 23% (P < 0.01; n = 8), regardless of pituitary cell type (Fig. 6, A–C). Furthermore, direct activation of phospholipasec by m-3M3FBS [2,4,6-trimethyl-N-C3-(trifluromethyl)phenylybenzesulfo-nomide] (50 μm) strongly attenuated I_h current in pituitary gonadotrophs (Fig. 6E). Based on these data, we concluded that I_h current is regulated by phosphoinositides in pituitary cells.

Fig. 6.

Regulation of HCN channels by phosphoinositides in gonadotrophs and somatotrophs. A–C, Inhibitory effects of wortmannin, a phosphatidylinositol-3 and phosphatidylinositol-4 kinase inhibitor, on I_h in pituitary cells. Representative traces from somatotrophs (A) and gonadotrophs (B) and mean values of amplitude of I_h and τ in both cell types (n = 8). D and E, Inhibition of I_h by 1 nm GnRH (D) and 50 μm m-3M3FBS [2,4,6-trimethyl-N-C3-(trifluoromethyl)phenylybenzesulfo-nomide], a phospholipase C activator (E), in gonadotrophs. s, Seconds; WT, wortmannin.

Discussion

It is well established that cyclic nucleotides can modulate electrical activity through cAMP/cGMP-dependent kinases that phosphorylate several channels, including the background K⁺ conductance in corticotrophs (30, 31), voltage-gated K⁺ channels (32), L- and T-type voltage-gated Ca²⁺ channels (33–35), tetrodotoxin-sensitive (36) and -resistant voltage-gated Na⁺ channels in somatotrophs (37), and transient receptor potential type C channels in all endocrine pituitary cells (38). Cyclic nucleotides could also modulate electrical activity in excitable cells by activation of CNG and HCN channels. Such a paradoxical role for these channels comes from their permeability properties (CNG are practically nonselective cation channels, and HCN channels are weakly K⁺-selective channels) and unique voltage dependence of HCN channels. Both second messengers contribute to regulation of CNG channels, with cAMP being more potent than cGMP in regulation of HCN channels (1).

Pituitary cells express a very sophisticated nitric oxide-synthase-soluble guanylyl cyclase signaling pathway (39–41). The mRNA transcripts for CNG channels were also found in a mixed population of pituitary cells (18). Should such transcripts be translated to their protein products, activation of these channels should stimulate firing of AP. However, here we show that the application of 8-Br-cGMP caused the reverse effect on the excitability of pituitary cells for these channels: it hyperpolarized the plasma membrane and decreased the firing frequency, in contrast to cAMP, which depolarized the plasma membrane and stimulated firing of AP. The inhibitory effect of cGMP on electrical activity is consistent with a role of protein kinase G in the regulation of BK-type Ca²⁺-controlled K⁺ channels (42). These channels are expressed in all endocrine pituitary cells, and their activation leads to inhibition of spontaneous electrical activity (43). Thus, cGMP plays an opposite role in the control of excitability of endocrine pituitary cells compared with cAMP, which is consistent with the inhibitory effect of nitric oxide signaling pathway on hormone release by cultured pituitary cells (44–46). Effects of NO/cGMP in intact pituitary gland are not only related to the electrical status of endocrine cells, but also may affect pituitary blood flow rates and oxygen supply by capillaries, as well as oxygen consumption (47, 48).

Previous studies have also described the expression of the HCN1 subunit and I_h in immortalized AtT-20 pituitary cells (tumor-derived mouse corticotrophs), which were weakly modulated by forskolin (16). The I_h was also identified in immortalized rat lactosomatotroph GH₃ cells and cultured rat somatotrophs (13, 14) but forskolin-induced changes in the properties of the I_h current were not observed. Here, we show that transcripts for all four HCN subunits are expressed in anterior pituitary cells. The protein expression of HCN2, HCN3, and HCN4 in pituitary cells and GH₃ lactosomatotrophs was confirmed using Western blot analysis. The apparent molecular weights corresponded to those previously reported for HCN subunits (Refs. 20 and 21 and references within). HCN2 and HCN3 appeared as doublets, probably due to posttranslational glycosylation (21). Protein expression of HCN1 could not be demonstrated, consistent with the low level of HCN1-mRNA transcript product and might reflect a cell type-specific expression pattern of HCN subunits in endocrine pituitary cells.

Furthermore, three lines of evidence suggested that pituitary lactotrophs, somatotrophs, thyrotrophs, and gonadotrophs, as well as other unidentified endocrine pituitary cells, express functional HCN channels. First, pituitary cells developed a slow-activating inward current upon hyperpolarization to a value of −60 mV or more negative. Second, the biophysical properties of this current (activation of this current by hyperpolarizing voltages, time course of activation, lack of inactivation during sustained recording, rapid deactivation, and activation curve obtained by tail current analysis) are consistent with the expression of these channels. Third, pharmacological profile of this current (reversible inhibition by 1 mm Cs⁺ and slow and irreversible inhibition by ZD7288) also supports the functional expression of HCN channels in endocrine pituitary cells, as was shown previously in other neuroendocrine cell types (49).

The majority of gonadotrophs, thyrotrophs, and somatotrophs express HCN channels. On the other hand, these channels were present in only a fraction of lactotrophs, which is consistent with the functional heterogeneity of these prolactin-producing cells (50, 51). What is the role of these channels in pituitary cell functions? Although the current density was relatively low in both normal and immortalized pituitary cells, such a current may still profoundly influence the resting membrane potential because pituitary cells are small and the input resistance of these cells is usually more than 5 GΩ. The contribution of the I_h current to the resting membrane conductance under normal ion concentrations was confirmed in experiments with hyperpolarizing current injection in gonadotrophs. In these cells, the voltage response to a hyperpolarizing current pulse showed slowly developing inward rectification, which was blocked by extracellular Cs⁺ and was absent in cells lacking I_h. This could indicate that I_h currents in pituitary cells may contribute to set the resting membrane potential in the range where other channels could account for the slow depolarization. It is important to stress that our experiments were done in isolated pituitary cells in vitro, and that in vivo these cells are organized as a large-scale network (52). The existence of such network, however, does not argue against the relevance of voltage-gated Ca²⁺ influx in rapid hormone release (48). Thus, it is reasonable to speculate that the operation of HCN channels is also preserved in interconnected cells.

Based on experiments with MDL12330A, an inhibitor of adenylyl cyclase, we also concluded that the basal cyclic nucleotide production in the majority of pituitary cells in vitro is high enough to partially or fully activate the I_h current. A similar effect was also observed using MDL12330A in GH₃ cells and was overcome by the subsequent application of forskolin (14). This also explains the relatively weak effect of forskolin on I_h activation. We also frequently observed a decrease in the AP frequency in spontaneously firing cells with inhibited I_h, indicating that these channels contribute to the electrical activity in a manner comparable to that observed in GnRH neurons (49). In none of our experiments we observed abolition of spontaneous electrical activity, further indicating the role of other channels in slow depolarization. Among others, the background Na⁺ conductance is present in all endocrine pituitary cells (53–55) and contributes to spontaneous electrical activity and accompanied Ca²⁺ transients (38). Basal cAMP production is down-regulated in vivo by several hypothalamic and intrapituitary factors, including dopamine and somatostatin (1), which suggest that facilitation of adenylyl cyclase activity in physiological conditions should lead to activation of HCN channels and firing of AP.

The activity of these channels is not only determined by the status of adenylyl cyclase activity, but also affected by G_q/11-coupled receptors signaling through phospholipase C pathway. Here we show that pharmacological activation of this enzyme leads to inhibition of I_h in gonadotrophs. Activation of GnRH receptors also causes inhibition of I_h. Experiments with wortmannin further indicated the potential role of PIP₂ in the regulation of these channels in gonadotrophs. This is consistent with numerous reports about the requirement of PIP₂ for HCN function (8, 56). Thus, it is reasonable to suggest that the G_q/11-coupled receptor-mediated shift from spontaneous electrical activity and Ca²⁺ influx to Ca²⁺ mobilization from intracellular pools is accompanied by a silencing of electrical activity mediated not only by the activation of apamin-sensitive K⁺ channels (23, 24) but also by the inhibition of HCN channels.

In conclusion, here we show that cAMP and cGMP exhibit opposite effects on the excitability of isolated endocrine pituitary cells; cGMP inhibits whereas cAMP stimulates electrical activity. We further show that the majority of pituitary gonadotrophs, thyrotrophs, somatotrophs, and a fraction of lactotrophs express HCN channels with biophysical, pharmacological, and immunoreactive properties comparable to those observed in other excitable cells expressing HCN channels. In cultured cells, basal adenylyl cyclase activity is sufficient to activate these channels, leading to facilitation of spontaneous firing of AP, and further stimulation of this enzyme enhances HCN channel activity. In contrast, activation of phospholipase C inhibits I_h current, presumably by depleting plasma membrane PIP₂ levels. In vivo, these channels may contribute to facilitated electrical activity and enhanced hormone release by activated G_s-coupled receptors and may play an important role in restoration of electrical activity in lactotrophs and somatotrophs after removal of their respective inhibitory cues, dopamine and somatostatin.

Materials and Methods

Chemicals

TRH, GnRH, GHRH, and somatostatin were purchased from Bachem (Torrance, CA), MDL12330A and 3,3′-(hydroxynitrosohydrazino) bis-1-propanamine from Alexis Biochemicals (San Diego, CA), and apomorphine, bromocryptine mesylate, tetrodotoxin, and ZD7288 were obtained from Tocris (Ellisville, MO). All other drugs and chemicals were purchased from Sigma (St. Louis, MO).

Cell cultures

Experiments were performed on anterior pituitary cells from normal postpubertal female Sprague Dawley rats obtained from Taconic Farms (Germantown, NY). Euthanasia was performed by asphyxiation with CO₂, and the anterior pituitary glands were removed after decapitation. Experiments were approved by the National Institute of Child Health and Human Development (NICHD) Animal Care and Use Committee. Anterior pituitary cells were mechanically dispersed after treatment with trypsin and cultured as mixed cells or enriched fractions in medium 199 containing Earle's salts, sodium bicarbonate, 10% heat-inactivated horse serum, penicillin (100 U/ml), and streptomycin (100 μg/ml), as described previously (57).

RT-PCR analysis of HCN isoform expression

Analysis of relative gene expression was done using real-time quantitative PCR and the comparative Ct method (19). For this purpose the predesigned Taq-Man Gene Expression Assays with LightCycler TaqMan Master mix (Applied Biosystems, Foster City, CA) and Lightcycler 2.0 Real-time PCR system (Roche, Indianapolis, IN) were used. To compare the relative expression levels of individual HCN channels, the levels were calibrated against HCN2 (set to 100%). Applied Biosystems predesigned Taq-Man Gene Expression Assays were used: for HCN1, Rn00584498_m1; HCN2, Rn01408572_mH; HCN3, Rn00586666_m1; HCN4, Rn00572232_m1; and GAPDH, Rn01462662_g1.

Western blot analysis of HCN channel isoforms

Primary pituitary cells or GH₃ cells in culture were lysed in radioimmunoprecipitation assay buffer containing 50 mm Tris HCl (pH 7.4), 150 mm NaCl, 1% Nonider P-40, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate, and supplemented with protease inhibitors. Samples were separated by Tris-glycine SDS-PAGE and transferred onto polyvinylidene difluoride membranes. The membrane was blocked for 1 h at room temperature and then incubated overnight at 4 C with one of the primary antibodies: anti-HCN2 [Abcam, Inc., Cambridge, MA; antibody (ab)19346], anti-HCN3 (Abcam, ab109807), or anti-HCN4 (Abcam, ab69054), all diluted 1:1000, or anti-β-tubulin I antibody (Sigma Aldrich, T7816) diluted 1:10000. All incubations were performed in SuperBlock Blocking Buffer (Thermo Fisher Scientific. Pittsburgh, PA) supplemented with 0.1% Tween-20. After incubation with peroxidase-conjugated goat antirabbit (HCN2 and -4) or donkey antigoat secondary antibody (HCN3) diluted 1:10000 (Santa Cruz Biotechnology, Inc., Santa Cruz, CA), blots were incubated with SuperSignal West Pico Chemiluminescent Substrate Kit (Thermo Fisher Scientific), and bands were visualized on FluorChem E Digital imaging System (ProteinSimple, San Jose, CA).

Cyclic nucleotide measurements

Cyclic nucleotide production was monitored using static cultures of anterior pituitary cells. Briefly, cells (1 million per well) were plated in 24-well plates in medium 199 and incubated overnight at 37 C under 5% CO₂-air and saturated humidity. The following day, medium was removed, cells were washed and then stimulated at 37 C under 5% CO₂-air and saturated humidity with 1 mm 3-isobutyl-1-methylxanthine. Cyclic nucleotides were measured in incubation medium by RIA using specific antisera provided by Albert Baukal (NICHD, Bethesda, MD). ¹²⁵I-labeled cAMP and ¹²⁵I-labeled cGMP tracers were purchased from PerkinElmer Life Sciences (Boston, MA).

Electrophysiological recordings

Membrane voltage potential and whole-cell currents were measured using amphotericine-perforated patch-clamp technique. All experiments were performed at room temperature. Cells cultured on 35-mm culture dishes (100,000 per dish) were washed twice before connecting to a relatively fast gravity-driven microperfusion system (ALA Scientific Instruments, Westbury, NY) with common outlet. Cells were continuously perfused with an extracellular solution containing: 150 mm NaCl, 3 mmKCl, 2 mm CaCl₂, 1 mm MgCl₂, 10 mm 4-(2-Hydroxyethyl)piperazine-1-ethanesulfonic acid, and 10 mm glucose starting from at least 10 min before recording at flow rate of approximately 2 ml/min. The pH was adjusted to 7.3 with NaOH. For voltage-clamp measurement of I_h currents, composition of extracellular solution was modified as follows: 113 mmNaCl, 20 mm KCl, and 20 mm tetraethylammonium chloride, 2 mmCaCl₂, 1 mm MgCl₂, 10 mm 4-(2-hydroxyethyl) piperazine-1-ethanesulfonic acid, 10 mm glucose, 1 mm 4-aminopyridine, 1 mm BaCl₂, 1 mm CdCl₂, and 0.001 mm tetrodotoxin. Patch pipettes were pulled from borosilicate glass (World Precision Instruments, Sarasota, FL) and heat polished to a tip resistance of 5–7 mΩ. Pipette solution contained: 90 mm K-aspartate, 50 mm KCl, 3 mm MgCl₂, and 10 mm 4-(2-hydroxyethyl)piperazine-1-ethanesulfonic acid, with pH adjusted to 7.2 with KOH. Before measurement, amphotericine B was added to the pipette solution from a stock solution (20 mg/ml in dimethyl sulfoxide) to obtain a final concentration of 200 μg/ml. Current-clamp and voltage-clamp recordings were performed using Axopatch 200B amplifier (Molecular Devices, Sunnyvale, CA). Recordings started when series resistance dropped below 100 mΩ for current-clamp or 40 mΩ for voltage-clamp recordings. Series resistance was compensated by more than 50%. Membrane potentials were corrected on-line for liquid junction potential of 9.9 mV. A difference of 1.9 mV between junction potential of control extracellular solution and modified 20 mm KCl solution used for I_h current recording was neglected. In some voltage-clamp experiments, P/4 protocol was applied on line to subtract leak conductance from current traces. Drugs dissolved to final concentration in extracellular solution were delivered to recording chamber by the same perfusion system, and less than 200 msec was required for exchange solutions around the patched cells, as described previously (58).

Calculations

The amplitude of I_h was calculated by subtracting the instantaneous current amplitude from the steady-state current at the end of the test pulse to −120 mV. The density of I_h was calculated by dividing the current amplitude by the membrane capacitance for each given cell. To characterize I_h activation kinetics, current traces elicited at −120 mV were fitted by a single exponential function or by the sum of two exponentials, and weighted tau (τ_w) was calculated as follows: τ_w = (τ1*A1+ τ2*A2)/(A1+A2), where A1 and A2 are the relative amplitudes of the two exponential components. Activation curves were determined from plots of tail current amplitudes (measured at −40 mV) as a function of test voltage during 7-sec hyperpolarizing steps. Normalized amplitudes of tail currents I/I_max were then plotted against test voltage and fitted with the Boltzmann equation: I/I_max = 1/{1+exp[(V_test − V_1/2)/s]}, where V_test is the hyperpolarizing step potential, V_1/2 is the half-activation potential, and s is the slope factor. Data values are expressed as mean ± sem. Statistical comparisons were made using the one-sample Student's t test. Significance was set at P < 0.05 or higher.