Paradoxical Role of Large-Conductance Calcium-Activated K⁺ (BK) Channels in Controlling Action Potential-Driven Ca²⁺ Entry in Anterior Pituitary Cells

Abstract

Activation of high-conductance Ca²⁺-activated K⁺ (BK) channels normally limits action potential duration and the associated voltage-gated Ca²⁺ entry by facilitating membrane repolarization. Here we report that BK channel activation in rat pituitary somatotrophs prolongs membrane depolarization, leading to the generation of plateau-bursting activity and facilitated Ca²⁺ entry. Such a paradoxical role of BK channels is determined by their rapid activation by domain Ca²⁺, which truncates the action potential amplitude and thereby limits the participation of delayed rectifying K⁺ channels during membrane repolarization. Conversely, pituitary gonadotrophs express relatively few BK channels and fire single spikes with a low capacity to promote Ca²⁺ entry, whereas an elevation in BK current expression in a gonadotroph model system leads to the generation of plateau-bursting activity and high-amplitude Ca²⁺transients.

Although all secretory anterior pituitary cells are of the same origin, they differ with respect to their Ca²⁺ signaling and secretory patterns (Stojilkovic and Catt, 1992; Kwiecien and Hammond, 1998). A majority of somatotrophs in vitro generate spontaneous high-amplitude fluctuations in intracellular Ca²⁺ concentration ([Ca²⁺]_i) that are abolished by the removal of extracellular Ca²⁺ or by the addition of Ca²⁺ channel agonists and antagonists (Lewis et al., 1988; Tomic et al., 1999). In parallel to Ca²⁺ signaling, growth hormone secretion from anterior pituitary cells in perifusion or static incubation experiments is high in the absence of any stimuli, and this basal secretion is inhibited by extracellular Ca²⁺ removal (Stojilkovic et al., 1988;Tomic et al., 1999). Gonadotrophs also exhibit extracellular Ca²⁺-dependent and dihydropyridine-sensitive [Ca²⁺]_ifluctuations; however, these [Ca²⁺]_ifluctuations are much smaller in amplitude than those observed in somatotrophs (Stojilkovic et al., 1992; Li et al., 1995; Kwiecien and Hammond, 1998). Moreover, basal-luteinizing hormone release is low and is not affected by the removal of extracellular Ca²⁺ or the inhibition of voltage-gated Ca²⁺ channels (VGCC; Stojilkovic et al., 1988). Such a difference in the Ca²⁺signaling and secretory patterns between somatotrophs and gonadotrophs is in accord with their modes of regulation by hypothalamic neurohormones. Growth hormone secretion is stimulated by growth hormone-releasing hormone (GHRH) and other known releasing factors and is inhibited by somatostatin and dopamine (Muller et al., 1999). In contrast, luteinizing hormone release is stimulated by gonadotropin-releasing hormone, whereas no known hypothalamic inhibitory factors have been identified (Sealfon et al., 1997).

The ionic and cellular mechanisms that endow somatotrophs, but not gonadotrophs, with the ability to generate high-amplitude [Ca²⁺]_i transients that are sufficient to trigger exocytosis in the absence of hypothalamic or local control are not known. Here we examined the patterns of action potential (AP) firing and the underlying [Ca²⁺]_i signals in female rat somatotrophs and gonadotrophs by simultaneous measurements of membrane potential and [Ca²⁺]_i in perforated patch-clamped cells. Our results revealed that, although a majority of somatotrophs and gonadotrophs exhibit spontaneous AP firing, there are distinct differences in the profile of the AP waveform and its capacity to drive Ca²⁺entry between the two cell types. In somatotrophs, plateau-bursting activity accounts for the generation of the high-amplitude [Ca²⁺]_itransients, whereas single spiking accounts for the low-amplitude [Ca²⁺]_i transients in gonadotrophs. In an extensive search for the underlying ionic currents mediating the discrete patterns of AP-driven Ca²⁺ entry and secretion in these two anterior pituitary cell types, our experiments revealed selective expression of the large-conductance calcium-activated K⁺ (BK) channels in somatotrophs. Furthermore, in contrast to the typical negative feedback role of BK channels in controlling voltage-gated Ca²⁺influx observed in other cell types (Kaczorowski et al., 1996; Sah, 1996; Vergara et al., 1998), these channels in somatotrophs act as positive feedback regulators of AP-driven Ca²⁺ entry by promoting the generation of the plateau bursting. Finally, we addressed the mechanism for such a paradoxical role of BK channels in controlling AP-driven Ca²⁺ entry in somatotrophs.

MATERIALS AND METHODS

Pituitary cell culture and cell identification.Anterior pituitary glands were excised from adult female Sprague Dawley rats (Taconic Farms, Germantown, NY) and dispersed into single cells by using a trypsin/DNase (Sigma, St. Louis, MO) cell dispersion procedure as described previously (Stojilkovic et al., 1988). Enriched somatotroph populations were obtained via a discontinuous Percoll density-gradient cell separation procedure as described previously (Koshimizu et al., 2000). Somatotrophs were identified further by their cell type-specific morphology and responses to the known neuroendocrine modulators GHRH and somatostatin. Gonadotrophs were identified initially by their cell type-specific morphology and, subsequent to experimentation, by the addition of gonadotropin-releasing hormone, which stimulates small-conductance Ca²⁺-activated (SK) K⁺ current and [Ca²⁺]_ioscillations only in gonadotrophs (Stojilkovic et al., 1992; Tse and Hille, 1992).

Electrophysiology. Current- and voltage-clamp recordings were performed at room temperature with an Axopatch 200B patch-clamp amplifier (Axon Instruments, Foster City, CA) and were low-pass filtered at 2 kHz. Unless otherwise indicated, all ionic currents and membrane potentials were measured with the perforated patch recording technique (Rae et al., 1991). Briefly, an amphotericin B (Sigma) stock solution (60 mg/ml) was prepared in DMSO and stored for up to 1 week at −20°C. Just before use the stock solution was diluted in pipette solution and sonicated for 30 sec to yield a final amphotericin B concentration of 240 μg/ml. Patch electrodes used for perforated patch recordings were fabricated from borosilicate glass (1.5 mm outer diameter; World Precision Instruments, Sarasota, FL) with a Flaming Brown horizontal puller (P-87, Sutter Instruments, Novato, CA). Electrodes were heat-polished to a final tip resistance of 3–6 MΩ and then coated with Sylgard (Dow Corning, Midland, MI) to reduce pipette capacitance. Pipette tips were immersed briefly in amphotericin B-free solution and then backfilled with the amphotericin B-containing solution. A series resistance of <15 MΩ was reached 10 min after the formation of a gigaohm seal (seal resistance, >5 GΩ) and remained stable for up to 1 hr. When necessary, series resistance compensation was optimized. All current recordings were corrected for linear leakage and capacitance by using a P/-N procedure. An average membrane capacitance (C_m) of 4.6 ± 0.2 and 7.5 ± 0.2 pF was recorded in somatotrophs and gonadotrophs, respectively. Pulse generation, data acquisition, and analysis were done with a PC equipped with a Digidata 1200 analog-to-digital (A/D) interface in conjunction with Clampex 8 (Axon Instruments). All values in the text are reported as mean ± SEM. Differences between groups were considered to be significant when p < 0.05, using the paired Student's t test.

Simultaneous recording of [Ca²⁺]_i and membrane potential or current. Pituitary cells were incubated for 15 min at 37°C in phenol red-free medium 199 containing Hanks' salts, 20 mm sodium bicarbonate, 20 mm HEPES, and 0.5 μm indo-1 AM (Molecular Probes, Eugene, OR). Membrane potential or ionic current was recorded as described above, and bulk [Ca²⁺]_iwas monitored simultaneously by a Nikon photon counter system as described previously (Van Goor et al., 2000). The membrane potential or current and bulk [Ca²⁺]_i were captured simultaneously at the rate of 5 kHz, using a PC that was equipped with a Digidata 1200 A/D interface in conjunction with Clampex 8 (Axon Instruments). The [Ca²⁺]_i was calibrated in vivo according to Kao (1994), and the values for R_min,R_max,S_f,480/S_b,480, and K_D were determined to be 0.75, 3.40, 2.45, and 230 nm, respectively. In some cases the net change in [Ca²⁺]_i was reported and was determined by subtracting the baseline [Ca²⁺]_i from the peak [Ca²⁺]_i that was reached during the spike waveform or command potential.

Chemicals and solutions. For the recording of electrical membrane activity and total ionic current, the extracellular medium contained (in mm): 120 NaCl, 2 CaCl₂, 2 MgCl₂, 4.7 KCl, 0.7 MgSO₄, 10 glucose, and 10 HEPES, pH-adjusted to 7.4 with NaOH, and the pipette solution contained (in mm): 50 KCl, 90 K⁺-aspartate, 1 MgCl₂, and 10 HEPES, pH-adjusted to 7.2 with KOH. So that the delayed rectifying K⁺ current (I_DR) could be isolated, the bath contained paxilline, apamin [to block the large- and small-conductance (SK) calcium-activated K⁺ currents (I_KCa)], and tetrodotoxin (to block voltage-gated Na⁺ currents). To isolate voltage-dependent Ca²⁺ currents (I_Ca) or to introduce exogenous Ca²⁺ buffers into the cytosol, we used conventional whole-cell recording techniques. For isolated calcium current recordings, 20 mm TEA and 0.001 mm tetrodotoxin were added to the extracellular medium, and the pipette contained (in mm): 120 CsCl, 20 TEA-Cl, 4 MgCl₂, 10 EGTA, 9 glucose, 20 HEPES, 0.3 Tris-GTP, 4 Mg-ATP, 14 CrPO₄, and 50 U/ml creatine phosphokinase, pH-adjusted to 7.2 with Tris base. For the introduction of exogenous Ca²⁺ buffers into the cytoplasm, 20 mm NaCl and 100 nm apamin were added to the extracellular medium, and the pipette contained (in mm): 130 K⁺-aspartate, 10 KCl, 1 MgCl₂, 9 glucose, 20 HEPES, 0.3 Tris-GTP, 4 Mg-ATP, 14 CrPO₄, 50 U/ml creatine phosphokinase, and 0.1 EGTA or 0.1 BAPTA, pH-adjusted to 7.2 with KOH. Under these whole-cell recording conditions there was no “rundown” inI_Ca (data not shown). All reported membrane potentials and ionic currents were corrected on-line for a liquid junction potential of +10 mV between the pipette and bath solution, except for I_Ca, which required no correction (Barry, 1994). The bath contained <500 μl of saline and was perifused continuously at a rate of 2 ml/min via a gravity-driven superfusion system. Stock solutions of iberiotoxin and apamin were prepared in double-distilled deionized water, whereas stock solutions of paxilline and BAPTA AM were prepared in DMSO. All chemicals were obtained from Sigma and Aldrich (Milwaukee, WI).

Gonadotroph model cell. To simulate the expression of BK channels and their impact on the pattern of AP firing in pituitary cells, we used the previously developed mechanistic biophysical model of gonadotrophs (Li et al., 1995). This gonadotroph model cell was based on experimentally derived electrical membrane and ionic channel properties and included the following conductances: a leak current, two types of Ca²⁺ channels (an L-type and a T-type), and two types of K⁺ channels (the delayed rectifier and the SK-type Ca²⁺-activated K⁺ channel). Because SK K⁺ channels are activated by Ca²⁺ inside the cell, Ca²⁺ handling in these cells was modeled by taking into account Ca²⁺ entry through VGCC, Ca²⁺ extrusion by plasma membrane Ca²⁺ pumps, and Ca²⁺ release and uptake by a passive intracellular Ca²⁺ store. The original model did not include a BK channel and exhibited tonic spiking activity similar to that observed in native gonadotrophs (for details, see Li et al., 1995).

In the present study the impact of BK channel expression on the pattern of AP firing in gonadotrophs was examined by incorporating a BK-type K⁺ channel into the original gonadotroph cell model. To do this, we modified the original gonadotroph cell model in two ways: (1) a BK-type Ca²⁺-activated K⁺ current (I_BK) was incorporated into the voltage equation, and (2) the cytosolic space was divided into two interconnected subcompartments: a submembrane shell and an interior core region. To simulate the changes in [Ca²⁺]_i occurring near the open pore of the VGCC (domain Ca²⁺; C_d) in the gonadotroph model cell, we modeled the profile of Ca²⁺ distribution in a buffered medium near a point source of Ca²⁺ flux. The concentration profiles can be approximated by hemispherically symmetric steady-state solutions to partial differential equations that describe the binding of Ca²⁺ to buffer molecules and the diffusion of Ca²⁺ and their buffers (Neher, 1998; Smith et al., 2001). We used the simplest distribution formula obtained with an excess buffer approximation (for details, see Smith et al., 2001):

$$C(r)=C_{\text{s}}+\varsigma e^{-\text{r}/\lambda }/(2\pi \text{Dr}).$$

Here, r is the distance from the flux source, ς is the source strength, D is the diffusion rate of Ca²⁺ in the medium,C_s is the background Ca²⁺ concentration in the submembrane shell, and:

$$\lambda =(D/(k^{+}[B]){)}^{1/2},$$

is the space constant that depends on the binding constant between Ca²⁺ and the bufferk⁺, the buffer concentration[B], and the Ca²⁺ diffusion rate D. We assume that the average distance between the L-type channel opening and the Ca²⁺binding site of the BK channel is 50 nm. For D = 250 μm/sec, k⁺ = 500 μm⁻¹/sec, and [B] = 200 μm, this distance is identical to the characteristic space constant λ. To model the domain Ca²⁺ concentration at the pore of an L-type Ca²⁺ channel, we made the source strength ς equal to the average single channel flux strength:

$$\varsigma =-i_{\text{Ca}-L}/2F.$$

(F is Faraday's constant). Therefore, the domain Ca²⁺ concentration:

$$C_{\text{d}}=C_{\text{s}}-I_{\text{Ca}-L}/(4e\pi NDF\lambda ),$$

where I_Ca−L is the L-typeI_Ca in the whole cell and Nis the number of L-type Ca²⁺ channels. In the model simulation we need only to specify the value of the lumped parameter p = 1/(4eπNDFλ) such that:

$$C_{\text{d}}=C_{\text{s}}-{PI}_{\text{Ca}-L}.$$

Implicit in this equation is the assumption that the Ca²⁺ microdomain appears and disappears instantaneously as the L-type channels open and close.

In simulations of BK channel expression into gonadotrophs, theI_BK was represented in the modified gonadotroph cell model by:

$$I_{BK}=g_{\text{k}-\text{bk}}{b}_{\infty }(V,C_{\text{d}}){\phi }_K,$$

where:

$${\phi }_K=V[1-exp(z_{K}F(V-V_K)/RT)]/(1-exp(z_{K}FV/RT)),$$

is the Goldman–Hodgkin–Katz driving force for K⁺ channels. The gating variableb_∞ is a function of both the membrane potential V and the domain Ca²⁺ concentration at the opening of L-type Ca²⁺ channels. Because gating is fast, we assumed that the dependence ofI_BK on V andC_d is instantaneous. This gating variable b_∞ is modeled by the formula:

$$b_{\infty }(V,C_{\text{d}})=1/[1+exp((V_{\text{b}}-V)/k_{\text{b}})],$$

where:

$$V_{\text{b}}(C_{\text{d}})=52(7-C_{\text{d}})(0.12+C_{\text{d}})/(9+C_{\text{d}})/(0.02+C_{\text{d}}),$$

and:

$$k_{\text{b}}(C_{\text{d}})=8(6+3C_{\text{d}})/(6+C_{\text{d}}),$$

with C_d in unit μm. These equations were obtained by directly fitting these expressions with experimental data on single BK channel properties in cultured rat muscle cells (Barrett et al., 1982).

In the revised model we divide the cytosolic space into a thin submembrane shell region and a spherical interior core, between which there are no barriers. Although Ca²⁺ can diffuse freely between these two compartments, they are treated as two distinct compartments to retain the differential features of Ca²⁺ signals in these two regions. The intracellular Ca²⁺ store is modeled as a distributed, passive, and linear Ca²⁺storage space that is capable of taking up and releasing Ca²⁺ rapidly. We also assumed that the Ca²⁺ level in the store is in constant equilibrium with the bulk [Ca²⁺]_i, indicating that the store behaves like a linear Ca²⁺ buffer that absorbs a fixed fraction of total intracellular Ca²⁺ at any fixed time. This reduced the number of Ca²⁺variables to two: the submembrane concentration (C_s) and the total intracellular concentration (C_T). Other Ca²⁺ variables such as the bulk Ca²⁺ concentration in the core (C_b), and the Ca²⁺ concentration in the endoplasmic reticulum Ca²⁺ stores (C_er) can be expressed as a function of C_s andC_T.

The model involves seven variables: membrane potential V, four gating variables, the submembrane Ca²⁺ concentration (C_s), and the total intracellular Ca²⁺ concentration measured by the shell volume:

$$C_T=C_{\text{s}}+{\varsigma }_1{C}_{\text{er}}+{\varsigma }_2{C}_{\text{b}},$$

where C_er andC_b are the concentrations in the endoplasmic reticulum (ER) Ca²⁺ store and the core, and ς₁ =V^e_er/V^e_shand ς₂ =V^e_bk/V^e_share the ratios between the effective volumes of the ER and the core to that of the submembrane shell. The effective volume refers to the physical volume of the compartment multiplied by the respective buffering capacity. Typically, the effective cytosolic volume is ∼100 times the physical volume because only ∼1 of Ca²⁺ 100 ions is free in the cytosol.

$$-C_{\text{m}}\text{dV}/\text{dt}=(g_{\text{ca}­\text{l}}{l}^2+g_{\text{ca}­\text{t}}{m}^{2}h){\phi }_{\text{Ca}}$$ Equation 1

$$+(g_{\text{k}­\text{dr}}n+g_{\text{k}­\text{sk}}{s}_{\infty }+g_{\text{k}­\text{bk}}{b}_{\infty })\phi K+g_{\text{l}}(V-V_L)$$

$$V_{\text{sh}}^{\text{e}}{\text{dC}}_{\text{s}}/\text{dt}=[C_{\text{eq}}(C_T)-C_{\text{s}}]/{\tau }_{\text{c}}-\alpha I_{\text{Ca}}-p(C_{\text{s}})$$ Equation 2

$$V_{\text{sh}}^{\text{e}}{C}_T/\text{dt}=-\alpha I_{\text{Ca}}-p(C_{\text{s}})$$ Equation 3

$$\text{dq}/\text{dt}=(q_{\infty }-q)/{\tau }_{\text{q}}(q \text{for} l,m,n,h),$$ Equation 4

where:

$${\phi }_I=V[1-exp(z_{\text{i}}F(V-V_{\text{i}})/RT)]/(1-exp(z_{\text{i}}FV/RT)),$$

where i for Ca²⁺ and K⁺ is the Goldman–Hodgkin–Katz driving force for ion i. Notice that the effect of varying domain Ca²⁺ concentration on φ_Ca is ignored for simplicity. For the activation gates l, m, n,

$$q_{\infty }=1/[1+exp((V_{\text{q}}-V)/k_{\text{q}})],$$

whereas for the inactivation gate h,

$$h_{\infty }=1/[1+exp((V-V_{\text{h}})/k_{\text{h}})].$$

The voltage-dependent time constants are: τ_l = τ_0l · τ(V) and τ_m = τ_0m · τ(V), with:

$$\tau (V)=exp((V_{\tau }-V)/k_{\tau })/[1+2exp(3(V_{\tau }-V)/k_{\tau })].$$

Then,

$$s_{\infty }(C_{\text{s}})=C_{\text{s}}^4/(C_{\text{s}}^4+K_{\text{s}}^4),$$

and b_∞(V,C_d) is given in the paragraph in which the BK channel is discussed. Also,

$$I_{\text{ca}}=I_{\text{CaL}}+I_{\text{caT}}=(g_{\text{ca}-1}{l}^2+g_{\text{ca}-\text{t}}{m}^{2}h){\phi }_{\text{Ca}},$$

$$A=1/({\varsigma }_1+{\varsigma }_2(1+k_{\text{e}}/L)),$$

$${\tau }_{\text{c}}={\kappa }^{-1}/(1+A),$$

$$C_{\text{eq}}(C_T)=C_{T}A/(1+A),$$

$$C_{\text{b}}=A(C_T-C),$$

$$C_{ER}=(1+k_{\text{e}}/l)C_{\text{b}},$$

$$p(C_{\text{s}})s_{\infty }(C_{\text{s}})=C_{\text{s}}^4/(C_{\text{s}}^4+K_{\text{s}}^4),$$

where L is the ER membrane Ca²⁺ permeability,k_e is the linear SERCA pump rate, andp(C_s) is the plasma membrane Ca²⁺ pump rate.

Parameter values used in the simulations include the following: the diameter of the cell, d = 10 μm;V^e_sh = 0.026 nl, ς₁ = 1, ς₂ = 0.5, α = 0.00518 (μm · nl)/(pA · sec),F/RT = 0.0375 mV⁻¹, C_m= 0.00314 nF, v_p = 0.05 (μm · nl)/sec, p = 0.1 μm/pA. In mV,V_ca = 125,V_K = −85, V_L= −60, V_l = −20,V_m = −30,V_h = −50_,V_n = −5.1,V_τ = −60,k_l = 12,k_m = 9,k_h = 4,k_n = 12.5,k_τ = 22. In sec, τ_{0 l} = 0.0185, τ_0m = 0.01, τ_h = 0.015, τ_n = 0.0225. In μm,K_s = 0.68 andK_p = 0.15. In nl/sec,L = 0.031, κ = 3, andk_e = 5. In nS,g_ca−l = 16,g_ca−t = 5,g_ca−dr = 1,g_K−sk = 0.6, g_K−bk = 0.46,g_l = 0.03.

RESULTS

Cell type-specific patterns of AP firing in pituitary cells

The patterns of AP firing and the associated changes in bulk [Ca²⁺]_i were compared between rat gonadotrophs and somatotrophs by monitoring membrane potential and [Ca²⁺]_isimultaneously. In gonadotrophs, spontaneous AP firing was observed in 52% of the cells that were examined (Fig.1, left panels). In all spontaneously active gonadotrophs (n = 26) the electrical membrane activity was characterized by the firing of single spikes at a frequency of 0.7 ± 0.1 Hz (mean ± SEM). The spike upstroke was rapid and reached a peak amplitude of 11.2 ± 2.1 mV. Spike repolarization was also relatively rapid, limiting its duration at one-half amplitude to 43 ± 15 msec. The interspike interval was characterized by a slow pacemaker depolarization from a baseline potential of −52 ± 1 mV and culminated in the initiation of an another spike. Removal of extracellular Ca²⁺ (n = 3) or the addition of L-type voltage-gated Ca²⁺channel blockers (n = 4) abolished spiking in all of the gonadotrophs that were examined in the present study (data not shown) and in previous studies (Stojilkovic et al., 1992; Li et al., 1995).

Fig. 1.

Distinct patterns of AP firing and [Ca²⁺]_i signaling in rat gonadotrophs and somatotrophs. A, Simultaneous recording ofV_m and [Ca²⁺]_i in an identified gonadotroph (left) and somatotroph (right), using the perforated patch-clamp recording configuration in the current-clamp mode. B, Expanded time scale of the AP and associated [Ca²⁺]_i signal identified in each cell type by the asterisks in A. Representative tracings from 26 gonadotrophs and 34 somatotrophs are shown.

In somatotrophs, spontaneous AP firing was observed in 88% (n = 34) of the cells that were analyzed (Fig. 1,right panels). Unlike gonadotrophs, the electrical membrane activity was characterized by the rhythmic firing of slow-wave plateau potentials that depolarized the membrane potential from −54 ± 2 to −24 ± 1 mV. Superimposed on the plateau potentials were multiple small-amplitude fast spikes. The first spike depolarized the membrane potential to −6.4 ± 1.7 mV, and the spikes that followed progressively decreased in amplitude during the plateau potential (the ratio of the first spike amplitude to that of the last spike was ∼3:1). Together, the plateau potentials and the associated small-amplitude spikes made up a single burst of electrical activity (hereafter referred to as plateau bursting) with a duration of 1.3 ± 0.2 sec. After each burst a pacemaker potential slowly depolarized the membrane potential toward the threshold for the next burst, resulting in a relatively slow frequency of 0.26 ± 0.03 Hz. Extracellular Ca²⁺ removal (n = 10) or the addition of VGCC blockers (n = 4) abolished plateau-bursting activity in all of the cells that were examined (data not shown).

The cell type-specific AP waveforms had different capacities to drive extracellular Ca²⁺ entry (Fig. 1). The single-spiking activity in gonadotrophs had a low capacity to drive Ca²⁺ entry, resulting in low-amplitude [Ca²⁺]_i transients (net change in [Ca²⁺]_i, 30 ± 5 nm; n = 26). In contrast, the plateau-bursting activity in somatotrophs had a high capacity to drive extracellular Ca²⁺ entry, resulting in high-amplitude [Ca²⁺]_i transients (net change in [Ca²⁺]_i, 583 ± 60 nm; n = 34). Despite the higher frequency of AP firing in gonadotrophs, the pattern of AP firing in somatotrophs gave rise to a higher average [Ca²⁺]_i when compared with that in gonadotrophs (average [Ca²⁺]_i measured over a 2 min period, 542 ± 54 nm in somatotrophs vs 130 ± 18 nm in gonadotrophs). These results indicate that the differences between gonadotrophs and somatotrophs with respect to the pattern of [Ca²⁺]_i signaling is attributable to the profile of the underlying AP waveforms and not to the (in)ability of each cell type to generate spontaneous AP firing.

Differential expression of BK channels between gonadotrophs and somatotrophs

We next examined whether there are differences in the ionic conductances between gonadotrophs and somatotrophs that may account for the cell type-specific AP firing patterns and their associated [Ca²⁺]_i signals. Although a similar group of ionic channels was found in each cell type, there was a marked difference in BK channel expression between gonadotrophs and somatotrophs. This was confirmed in experiments in which the expression of BK channels was analyzed by a two-pulse protocol (Fig. 2). This protocol consisted of an initial membrane depolarization to −10 mV for 100 msec (holding potential, −90 mV) to activate VGCCs. After this Ca²⁺ influx step the membrane potential was stepped from −10 to +90 mV for 500 msec, during which the magnitude of peak K⁺ current (I_K) activation was evaluated (Fig.2A). Because +90 mV is near the reversal potential for Ca²⁺ under our experimental conditions, there was no net Ca²⁺ influx when the membrane potential was stepped directly to +90 mV from a holding potential of −90 mV (Fig. 2B, bottom trace; basal [Ca²⁺]_i vs [Ca²⁺]_i during the test pulse, 61 ± 18 vs 64 ± 19 nm;p > 0.05; n = 15). Conversely, the application of the two-pulse protocol evoked a significant and similar rise in [Ca²⁺]_i in both cells types (Fig. 2B, top trace; net change in [Ca²⁺]_i: gonadotrophs, 160 ± 18 nm; somatotrophs, 155 ± 28 nm; n = 20). This is consistent with the similarity in the sustained voltage-gated Ca²⁺ current (I_Ca) density at −10 mV (somatotrophs, 4.3 ± 0.7 pA/pF; n = 5; gonadotrophs, 5.1 ± 0.9 pA/pF; n = 5) and theI_Ca–voltage relation in both cell types (data not shown; n = 5).

Fig. 2.

Differential expression of BK channels between gonadotrophs and somatotrophs. A, A two-pulse protocol was used to monitor I_KCaactivation by voltage-gated Ca²⁺ entry. This protocol consisted of a 100 msec conditioning pulse to −10 mV to activated VGCCs, followed by a 500 msec test pulse to +90 mV, during which the peak I_K was monitored.B, Change in [Ca²⁺]_ievoked by two-pulse protocol and by the test pulse alone in gonadotrophs (left) and somatotrophs (right). C, Extracellular Ca²⁺ removal reduced I_Kevoked by the two-pulse protocol in gonadotrophs and somatotrophs.D, The net I_KCaactivated by the two-pulse protocol in gonadotrophs (n = 9) and somatotrophs (n = 15) was obtained by subtracting the current evoked in Ca²⁺-deficient medium from the control current.E, Application of 1 μm paxilline reducedI_K evoked by the two-pulse protocol in gonadotrophs and somatotrophs. F, The net paxilline-sensitive I_BK activated by the two-pulse protocol in gonadotrophs (n = 11) and somatotrophs (n = 15). The mean ± SEMs of the peak I_KCa andI_BK evoked during the test pulse are shown in D and F, respectively. The peakI_BK isolated by 100 nm IBTX subtraction in gonadotrophs (n = 5) and somatotrophs (n = 10) was 0.089 ± 0.012 and 0.521 ± 0.36 nA/pF, respectively. To account for differences in cell size between gonadotrophs and somatotrophs, we normalized all currents to the membrane capacitance of each cell that was examined.

In gonadotrophs and somatotrophs the removal of extracellular Ca²⁺ reduced theI_K evoked during the test pulse (Fig.2C), indicating the presence of a Ca²⁺-sensitiveI_K(I_KCa) in both cell types. Isolation of I_KCa by subtracting the extracellular Ca²⁺-dependent current from the total current indicated that its peak amplitude was much greater in somatotrophs than in gonadotrophs (Fig. 2D). Like extracellular Ca²⁺ removal, the application of the highly specific BK channel blockers, 1 μm paxilline (Knaus et al., 1994; Sanchez and McManus, 1996) (Fig. 2E) or 100 nm iberiotoxin (IBTX; Galvez et al., 1990;Giangiacomo et al., 1992) (data not shown), reduced theI_K evoked during the test pulse in both cell types. Extracellular Ca²⁺removal in the presence of either BK channel blocker did not reduceI_K further. In addition, isolation of the BK current (I_BK) by subtracting the paxilline-sensitive (Fig. 2F) or IBTX-sensitive (Fig. 2, legend) current from the total current indicated that, within each cell type, the peakI_BK amplitude was similar to that ofI_KCa. These results indicate thatI_BK is the predominantI_KCa activated by the two-pulse protocol in both cell types. Moreover, the magnitude ofI_BK activation by voltage-gated Ca²⁺ entry is much greater in somatotrophs than in gonadotrophs. This is likely attributable to differences in BK channel expression, because the increase in [Ca²⁺]_i evoked by the Ca²⁺ influx step, theI_Ca–voltage relation, andI_Ca density measured at −10 mV are similar between gonadotrophs and somatotrophs (see above).

Dependence of plateau bursting on BK channel activation

The differential expression of BK channels between somatotrophs and gonadotrophs suggests that these channels may underlie the cell type-specific patterns of AP-driven [Ca²⁺]_i signaling. To test this, we examined the role of BK channels in shaping the profile of the AP waveform and the associated [Ca²⁺]_i signals in gonadotrophs and somatotrophs. We first examined whether a rise in [Ca²⁺]_i is required for I_BK activation and the generation of the distinct AP-firing patterns in each cell type. To do this, we examined the effects of the membrane-permeable Ca²⁺ buffer, BAPTA AM, onI_BK activation and the pattern of AP firing. In cells preloaded with 20 μm BAPTA AM for 45 min at 37°C, the two-pulse protocol did not activateI_BK in somatotrophs (Fig.3A) or gonadotrophs (data not shown). In all five gonadotrophs that were examined, the pattern of AP firing was not affected by BAPTA AM (data not shown). In contrast, in somatotrophs preloaded with BAPTA AM, the profile of the AP waveform was shifted from plateau bursting to single spiking (Fig.3B; n = 5).

Fig. 3.

Dependence of the plateau-bursting activity in somatotrophs on the activation of BK channels. A,Left, Representative I_Ktracings evoked by the two-pulse protocol in somatotrophs preincubated with DMSO (Control) or BAPTA AM (20 μm) for 45 min at 37°C. A,Right, Mean ± SEM of the peakI_K evoked by the two-pulse protocol in somatotrophs preincubated with DMSO (n = 5), BAPTA AM (n = 5), and BAPTA AM in the absence of extracellular Ca²⁺ (n = 5).B, Left, Simultaneous measurement of the membrane potential and [Ca²⁺]_i in somatotrophs preincubated with DMSO (Control) or BAPTA AM. B, Right, Expanded time scale of AP and associated [Ca²⁺]_i signals identified in the left panel byb₁ (Control) andb₂ (BAPTA/AM). Representative tracings from five controls and BAPTA AM-loaded somatotrophs are shown. C, D, Left, Simultaneous measurement of the membrane potential and [Ca²⁺]_i in somatotrophs before and during (horizontal bar) the application of 100 nm IBTX (n = 14) or 1 μm paxilline (n = 4). C, D, Right, Expanded time scale of the AP and associated [Ca²⁺]_i signals before (c₁ andd₁) and during the application of IBTX (c₂) or paxilline (d₂).

Similarly, the application of 100 nm IBTX or paxilline during spontaneous AP firing in somatotrophs shifted the profile of the AP waveform from plateau bursting to single spiking, which resulted in a dramatic decrease in the amplitude of the [Ca²⁺]_i transients (Fig. 3C,D). Specifically, the peak spike amplitude was increased, and there was a marked increase in the magnitude of spike repolarization (Table 1). Consequently, the plateau-bursting potential was abolished, and the capacity of AP firing to drive Ca²⁺ entry was reduced (Table 1). In addition to the shift from phasic to tonic spiking, the application of the BK channel blockers depolarized the baseline potential from −54 ± 2 to −49 ± 2 mV (p < 0.05; n = 18) (Fig. 3). In gonadotrophs the application of 100 nm IBTX during spontaneous AP firing did not alter the amplitude (peak spike amplitude: control, 9.5 ± 3.2 mV vs IBTX, 10.4 ± 4.3 mV;p > 0.05; n = 4) or duration (spike duration at one-half amplitude: control, 10.5 ± 2.0 msec vs IBTX, 12.7 ± 2.8 msec; p > 0.05; n = 4) of the single spikes. These results indicate that BK channel activation is required for generating the sustained plateau potential in somatotrophs, which prolongs AP duration and facilitates extracellular Ca²⁺ entry.

Table 1.

Effects of 100 nm IBTX and 1 μmpaxilline on the profile of the AP waveform and the associated [Ca²⁺]_i transients in spontaneously active rat somatotrophs

	Control	IBTX (14)	Control	Paxilline (4)
Peak spike amplitude (mV)	−5.4  ± 1.2	7.3  ± 1.9*	−3.5  ± 1.3	12.9  ± 1.3*
Net spike repolarization (mV)	21.0  ± 1.6	46.4  ± 3.0*	26.5  ± 1.0	58.3  ± 2.5*
Net Δ[Ca2+]i(nm)	375.8  ± 97.2	46.0  ± 7.8*	438.0  ± 39.1	123.2  ± 27.2*

The net spike repolarization was measured from the peak spike amplitude to the baseline potential recorded during the interspike period. The number of trials for each drug are shown in parentheses, and the asterisks indicate a significant difference from control values (p < 0.05).

BK channels limit delayed rectifying K⁺channel activation

We next examined the ionic mechanisms underlying the apparent paradoxical role of BK channels in facilitating AP-driven Ca²⁺ entry in rat somatotrophs. One possibility is that the IBTX- or paxilline-induced depolarization of the baseline potential shifts the pattern of AP firing from plateau bursting to single spiking. In thalamocortical neurons, for example, sustained membrane depolarization by current injection and the ensuing inactivation of T-type VGCC shifts the pattern of AP firing from bursting to single spiking (McCormick and Pape, 1990). To test whether depolarization of the baseline potential itself is sufficient to shift the pattern of AP firing in spontaneously active somatotrophs, we depolarized the membrane potential in a nonreceptor- and receptor-mediated manner, using KCl and GHRH, respectively. In five separate somatotrophs the addition of 5 mm KCl (total KCl concentration, 9.7 mm) depolarized the baseline potential and increased the burst frequency from 0.19 ± 0.06 to 0.29 ± 0.05 Hz (p < 0.05) but did not change the profile of the AP waveform (Fig.4A,B). Similarly, membrane depolarization by 100 nm GHRH increased the frequency of bursting from 0.16 ± 0.03 to 0.42 ± 0.08 Hz (p < 0.05; n = 8) but did not induce tonic spiking (Fig. 4C,D).

Fig. 4.

Effects of KCl- and GHRH-induced membrane depolarization on the pattern of AP firing in somatotrophs.A, Simultaneous measurement of membrane potential and [Ca²⁺]_i before and during the application of 5 mm KCl (horizontal bar) in a spontaneously active somatotroph. B, Expanded time scale of the AP and associated [Ca²⁺]_isignals before (single asterisks) and during (double asterisks) KCl application. C, Simultaneous measurement of membrane potential and [Ca²⁺]_i before and during the application of 100 nm GHRH (horizontal bar) in a spontaneously active somatotroph. D, Expanded time scale of the AP and associated [Ca²⁺]_isignals before (single asterisks) and during (double asterisks) KCl application. Dashed lines indicate the baseline potential before the application of KCl or GHRH.

In addition to shifting the pattern of AP firing from plateau bursting to single spiking, the application of BK channel blockers increased the peak spike amplitude (Figs. 3C,D,5A, Table 1), indicating that BK channel activation truncates spike amplitude in rat somatotrophs. In general, the spike amplitude determines the magnitude of delayed rectifying (DR) channel activation, which in turn controls the rate and magnitude of membrane repolarization and thus the capacity of AP firing to drive Ca²⁺ entry (Van Goor et al., 2000). To test whether BK channels limit DR channel activation, we determined the peak I_DR evoked by different peak spike amplitudes from the current–voltage relation of the isolated I_DR (for details, see Materials and Methods). In response to 1 sec depolarizing steps from −80 to +20 mV (holding potential, −90 mV), a slowly activating and inactivating I_DR was observed (Fig.5B). The I_DR measured at early (0–25 msec) times activated at membrane potentials more depolarized than −30 mV (Fig. 5C, filled circles), whereas the I_DRmeasured at late (990–1000 msec) times was observed at membrane potentials more depolarized than −40 mV (Fig. 5C,open circles). The discrepancy in the apparent activation potentials of the early and late I_DRwas attributable to the relatively slow activation of this current in somatotrophs (Fig. 5D). Within the range of peak spike amplitudes reached during plateau bursting and paxilline- or IBTX-induced single spiking (−10 to +10 mV), there was a marked increase in the peak I_DR (≈14 pA/mV), as illustrated in Figure 5, A and C.

Fig. 5.

Relationship between spike amplitude and delayed rectifying K⁺ channel activation in somatotrophs. A, Simultaneous measurement of membrane potential and [Ca²⁺]_i before and during the application of 100 nm IBTX in a spontaneously active somatotroph. The dashed lines indicate the change in peak spike amplitude with and without (IBTX) BK channel activation. B, Representative tracing of the isolated delayed rectifying K⁺ current (I_DR) in response to 1 sec membrane potential steps from a holding potential of −90 mV to −90, −20, −10, 0, and 10 mV. C, Current–voltage relation of the early (0–25 msec; filled circles) and late (990–1000 msec; open circles) I_DRevoked by 1 sec membrane potential steps from −80 to 20 mV (holding potential, −90 mV; mean ± SEM; n = 6).D, Time constant (τ) of I_DRactivation during 1 sec membrane potential steps from a holding potential of −90 mV to −20, −10, 0, and 10 mV. The time constant of activation for the I_DR was best determined by a single exponential fit.

We next examined whether the membrane potential remained depolarized for a sufficient period of time during the relatively rapid single spike to activate significantly moreI_DR than during plateau-bursting activity. To do this, we monitored the magnitude and kinetics ofI_DR activation in response to the application of a single spike or plateau burst in somatotrophs by the AP clamp recording technique. Prerecorded APs from a spontaneously active gonadotroph (spike AP) and somatotroph (burst AP) were used as the command potential waveforms in the voltage-clamp recording mode (Fig. 6A). Under isolated I_DR recording conditions (for details, see Materials and Methods) both AP waveforms activatedI_DR in somatotrophs. However, the peakI_DR amplitude evoked by the spike AP was much higher than that evoked by the burst AP (Fig.6B). In response to the spike AP, theI_DR activated during the spike upstroke, reached peak amplitude during the repolarization phase, and returned to baseline levels during the interspike membrane potential (Fig. 6C, left panel). In response to the burst AP, the I_DR also activated during the upstroke of the initial spike and reached peak amplitude during the repolarization phase. After the initial spike, however, theI_DR did not return to the baseline levels that were observed before the AP waveform (Fig. 6C,right panel), resulting in sustained activation ofI_DR during the plateau potential. These results indicate that, despite the slow activation ofI_DR (Fig. 5D), the membrane potential during single spiking remained depolarized for a sufficient period of time to activate significantly moreI_DR than that evoked by the lower amplitude spikes during plateau-bursting activity.

Fig. 6.

Delayed rectifyingI_K evoked by prerecorded single spike and plateau-burst command potentials in somatotrophs. A, A single spike (Spike AP) and plateau-burst (Burst AP) AP waveform were prerecorded from a spontaneously active gonadotroph and somatotroph, respectively, and then used as the command potential under voltage-clamp recording conditions. B, Representative current trace of the isolatedI_DR evoked by the spike (holding potential, −50 mV) and burst (holding potential, −60 mV) AP waveforms in the same somatotroph. The mean ± SEM of the peakI_DR evoked by the spike (open bar) and burst (hatched bar) AP waveform (n = 8) is shown on the right.Asterisks denote significant differences (p < 0.01, paired t test).C, Expanded time scales of the spike (left) and burst (right) AP waveforms (dashed lines) and the evokedI_DR (solid line) shown inA. Note the different y-axis scales between the left and rightpanels.

On the basis of these results, an increase in spike amplitude caused by the inhibition of BK channels should activate significantly moreI_DR, which would facilitate spike repolarization and reduce AP-driven Ca²⁺entry. To test this, we estimated the total current contributing to spike repolarization before and during BK channel inhibition by IBTX by multiplying the rate of spike repolarization (dV_m/dt) by the membrane capacitance (C_m; 4.6 ± 0.2 pF). During plateau bursting the repolarization rate of the initial spike was 2.3 ± 0.2 mV/msec, which corresponds to a net current of 10.6 ± 1.1 pA. Application of 100 nm IBTX significantly (p < 0.01; n = 14) increased the spike repolarization rate to 3.9 ± 0.3 mV/msec and the net current to 17.9 ± 1.3 pA. Therefore, despite BK channel inhibition there was more net outward current contributing to spike repolarization during single spiking than during plateau bursting. Moreover, these results indicate that BK channel activation truncates the peak spike amplitude, which limits the magnitude of DR channel activation.

Rapid activation of BK channels by domain [Ca²⁺]_i

So that the peak spike amplitude can be truncated, which leads to the reduction in I_DR, BK channel activation must be rapid. To test this, we simultaneously monitoredI_BK and bulk [Ca²⁺]_i in response to a modified two-pulse protocol. This protocol consisted of a series of Ca²⁺ influx steps ranging in duration from 0 to 300 msec, each of which was followed by a 1 sec test pulse to +90 mV (holding potential, −90 mV). In addition, to monitor the slow decline in bulk [Ca²⁺]_i, we held the membrane potential at −90 mV for 10 sec after the termination of the test pulse (Fig. 7A). In response to incremental increases in the duration of the Ca²⁺ influx step, there was a progressive increase in both the peak I_BK (Fig.7B,D) and bulk [Ca²⁺]_i (Fig.7C,D).

Fig. 7.

Dissociation betweenI_BK and bulk [Ca²⁺]_i kinetics in somatotrophs. A series of Ca²⁺ influx steps ranging from 0 to 300 msec was given before the application of a single test pulse to +90 mV (A), during which theI_K (B) and bulk [Ca²⁺]_i (C) were monitored simultaneously. D, The mean ± SEM of the peak I_K (solid line) and [Ca²⁺]_i (dotted line) from 19 somatotrophs was plotted against the Ca²⁺influx step duration. E, Expanded time scale of the tracings in B and C showing both theI_K (solid lines) and change in bulk [Ca²⁺]_i (dotted lines) evoked by Ca²⁺ influx steps of 0 and 5 msec (F), 25 and 50 msec (G), and 100, 200, and 300 msec (H) in duration. For clarity, the two-step protocol is not shown to scale, and not all steps are labeled.

However, the expanded time scale of the current and [Ca²⁺]_i tracings indicated that there was a dissociation in the kinetics of theI_BK and bulk [Ca²⁺]_i profiles (Fig. 7E–H). Specifically, brief Ca²⁺ influx steps (<20 msec) did not change the bulk [Ca²⁺]_i but activated the I_BK. In addition,I_BK decreased rapidly when Ca²⁺ influx was terminated by stepping the membrane potential to +90 mV (Fig. 7F). Although there was a progressive increase in bothI_BK and bulk [Ca²⁺]_i in response to longer (≥ 25 msec) Ca²⁺influx steps, I_BK decreased during the slow rise in bulk [Ca²⁺]_i (Fig.7G,H). Because of the rapid association kinetics of Ca²⁺ binding to indo-1 (k_on = 5 × 10⁸m/sec; Jackson et al., 1987), it is unlikely that the slow rise in the reported [Ca²⁺]_i is attributable to the inability of the Ca²⁺indicator dye to follow the rise in bulk [Ca²⁺]_i. Moreover, because of the discrepancy in the kinetics of the bulk [Ca²⁺]_i andI_BK, these results suggest that at least a fraction of the BK channels in somatotrophs responds to the rapid, high-amplitude increase in submembrane [Ca²⁺]_i in the vicinity of the VGCC (heretofore referred to as the domain [Ca²⁺]_i).

To test whether the rapid activation of BK channels is attributable to their activation by domain [Ca²⁺]_i, we used two different calcium buffers, BAPTA and EGTA, which bind Ca²⁺ with a similar affinity but at different rates. Because BAPTA (k_on = 6 × 10⁸m/sec;Tsien, 1980) binds Ca²⁺ ∼100× faster than EGTA (k_on = 1.5 × 10⁶m/sec; Tsien, 1980; Adler et al., 1991), it should be more effective at buffering domain [Ca²⁺]_i. Consequently, it should also be more effective than EGTA at inhibiting BK channels that respond to fluctuations in domain [Ca²⁺]_i. A similar approach has been used to investigate the relationship between BK channel activity and domain [Ca²⁺]_i in other cell types (Roberts, 1993; Robitaille et al., 1993).

The exogenous Ca²⁺ buffers were introduced into the cytoplasm via the recording electrode by standard whole-cell recording techniques, whereas the endogenous Ca²⁺ buffers were preserved with the perforated patch recording technique. In the presence of the endogenous Ca²⁺ buffers of the cell, incremental steps in the duration of Ca²⁺ influx resulted in a progressive increase in the peakI_BK amplitude (Fig.8A,D). In the presence of the slow Ca²⁺ buffer, EGTA (100 μm), there was a similar increase in the peakI_BK in response to short Ca²⁺ influx steps (<25 msec). However, the peak I_BK amplitude evoked by longer Ca²⁺ influx steps (≥ 25 msec) was reduced compared with that in the presence of the endogenous Ca²⁺ buffers (Fig.8B,D). In contrast, the introduction of a similar concentration of the fast Ca²⁺ buffer, BAPTA (100 μm), into the cytoplasm markedly attenuated I_BK activation by both short and prolonged Ca²⁺ influx steps (Fig. 8C,D). These results confirm that at least a fraction of the BK channels in somatotrophs is colocalized with VGCCs and responds to the associated fluctuations in domain [Ca²⁺]_i.

Fig. 8.

Effects of endogenous and exogenous Ca²⁺ buffers on I_BK in somatotrophs. Representative current tracings evoked by the two-pulse protocol in the presence of the endogenous Ca²⁺buffers (A) or the slow and fast exogenous Ca²⁺ buffers, 100 μm EGTA (B) and 100 μm BAPTA (C), respectively. The endogenous Ca²⁺ buffers were preserved by using the perforated patch recording configuration, whereas the exogenous Ca²⁺ buffers were introduced into the cytoplasm by the recording pipette via standard whole-cell recording techniques.D, The mean ± SEM of the isolatedI_BK evoked by the two-pulse protocol in the presence of the endogenous Ca²⁺ buffers (n = 15) or the exogenous Ca²⁺buffers EGTA (n = 5) or BAPTA (n = 3).

Rapid deactivation of BK channels by the clearance of domain [Ca²⁺]_i

In addition to reducing DR channel activation, BK channels also must deactivate and/or inactivate rapidly to prevent full membrane repolarization and to allow for the generation of the sustained plateau potential. To test this, we first examined the deactivation and/or inactivation kinetics of the I_BK. The decrease in I_BK during the test pulse may be attributable to deactivation of the BK channels by the rapid clearance of domain [Ca²⁺]_i and/or voltage-dependent inactivation of these channels (Solaro and Lingle, 1992; Prakriya et al., 1996; Prakriya and Lingle, 2000). To determine whether the BK channels in somatotrophs inactivate, we used a single step protocol, which consisted of 1 sec depolarizing steps from a holding potential of −90 mV to potentials ranging from −80 to +70 mV (Fig. 9A). TheI_BK was isolated by using current subtraction studies as described above. Both the early (0–25 msec;open circles) and sustained (990–1000 msec; filled circles) I_BK activated at membrane potentials more depolarized than −40 mV and were dependent on the underlying I_Ca–voltage relation in these cells (Fig. 9B,C). The sustained increase in the magnitude of I_BK activation during the 1 sec depolarizing steps indicates that the BK channels in somatotrophs do not inactivate. Thus, the rapid decrease inI_BK during the relatively gradual rise in bulk [Ca²⁺]_i(Fig. 7G,H) is likely attributable to the deactivation of the channels by the clearance of domain [Ca²⁺]_i and not to BK channel inactivation.

Fig. 9.

Voltage-dependent inactivation and deactivation properties of I_BK in somatotrophs. A, To determine whether the BK channels in somatotrophs inactivate during sustained membrane depolarizations, we applied 1 sec depolarizing voltage steps from −90 to +70 mV (holding potential, −90 mV). B, Representative current traces of the isolated I_BK from nine somatotrophs. The isolated I_BK was obtained by subtracting the IBTX- or paxilline-sensitive current from the total current.C, Current–voltage relation of the early (open circles; 0–25 msec) and late (filled circles; 990–1000 msec) I_BK shown in B. D, The relationship between the clearance of domain Ca²⁺ andI_BK activation was monitored by using a modified two-pulse protocol, during which the membrane potential was stepped back to −90 mV for 0–300 msec before the application of the test pulse. E, Representative current traces evoked by the modified two-pulse protocol. F, The mean ± SEM (n = 5) of the peak I_Kevoked during the test pulse in the absence of a Ca²⁺ influx step and after an interstep interval of 0–300 msec in duration. The continuous line is a single exponential fit to the data.

To demonstrate directly that the decline inI_BK is attributable to deactivation, we monitored I_BK activation after allowing domain [Ca²⁺]_i to clear for variable time periods. To do this, we stepped the membrane potential to −10 mV (holding potential, −90 mV) for 5 msec, which was sufficient to activate I_BK but did not increase bulk [Ca²⁺]_isignificantly (Fig. 7F). Then the membrane potential was stepped back to −90 mV for a variable duration (interstep duration) before being stepped to the test potential of +90 mV, during which I_BK was monitored (Fig.9D). In addition to terminating Ca²⁺ influx through VGCCs, the interpulse step to −90 mV also should remove any steady-state inactivation of the BK channels. The magnitude of I_BKevoked during the test pulse then was compared with that evoked during the test pulse given immediately after the 5 msec Ca²⁺ influx step. As shown in Figure 9,E and F, I_BK was decreased significantly after only brief interstep durations, confirming that the decline in I_BK is attributable to deactivation of the channels by the rapid decrease in domain [Ca²⁺]_iafter the termination of Ca²⁺ influx through VGCCs.

To test whether the rapid deactivation ofI_BK by the clearance of domain Ca²⁺ prevents BK channels from fully repolarizing the membrane potential, we monitored the amplitude and kinetics of both I_DR andI_BK activation during plateau-bursting activity via the AP clamp technique (see above). Application of a prerecorded burst AP evoked an outward current, which was reduced by the application of 1 μm paxilline. To isolate the I_BK underlying the generation of the burst AP, we subtracted the current evoked in the presence of paxilline from the total current (Fig.10A). Under these conditions I_BK was observed during the AP waveform, but not during the interburst period (Fig.10A, bottom panel), indicating that BK channels do not contribute to the baseline potential in these cells. We next compared the magnitude and kinetics ofI_DR andI_BK activation during the initial small-amplitude spikes of plateau-bursting activity (Fig.10B). The I_BKrapidly activated during the spike upstroke and then rapidly deactivated during spike repolarization. In contrast, a fraction of theI_DR remained activated during the interspike intervals. These results further indicate thatI_BK rapidly activates during the spike upstroke, which limits spike amplitude and prevents full activation ofI_DR (Fig. 6B). Moreover, I_BK also rapidly deactivates during the spike repolarization phase, which prevents full spike repolarization and allows for the generation of the plateau potential and prolonged Ca²⁺ influx.

Fig. 10.

Profile of theI_DR and I_BKunderlying the generation of plateau-bursting activity in somatotrophs.A, In seven somatotrophs theI_BK evoked by the prerecorded burst AP (top panel) was isolated from theI_DR by subtracting the paxilline-sensitive current from the total current. Representative current tracings are shown. B, Expanded time scale of the burst AP command potential (top panel) and the underlyingI_DR and I_BK shown in A. C, Expanded time scale of the burst AP waveform and the isolated I_DR andI_BK from the gonadotroph model cell expressing BK channels (Fig. 11C, right panel, top and bottom traces).

Addition of BK channels to a gonadotroph model cell converts single spiking to plateau bursting

Our experimental results indicate that the rapid activation/deactivation of BK channels by domain [Ca²⁺]_i is required for the generation of plateau bursting and the associated high-amplitude [Ca²⁺]_i transients in somatotrophs. To test this hypothesis further, we examined the impact of increasing BK channel expression in gonadotrophs. Unlike somatotrophs, gonadotrophs express few BK channels and exhibit single-spiking activity that has a low capacity to drive Ca²⁺ entry (Figs. 1, 2). These features were well captured by a mechanistic biophysical model of rat gonadotrophs (Fig.11A,B, left panels) previously developed by Li et al. (1995). Spontaneous AP firing in the model is described by a Hodgkin–Huxley-like set of equations, the parameters for which were derived from experimental data obtained from rat gonadotrophs. The major ionic channels contributing AP generation in gonadotrophs include T-type Ca²⁺ channels, L-type Ca²⁺ channels, DR channels, SK channels, and a small leak current (for details, see Materials and Methods).

Fig. 11.

Introduction of BK channels into the gonadotroph model cell shifts the pattern of AP firing from single spiking to plateau bursting. A, Pattern of AP firing and the associated changes in bulk Ca²⁺ concentration in the gonadotroph model cell in the absence (left) and presence (right) of BK channels. On the basis of the experimental evidence, BK channels in the model cells responded to fluctuations of domain Ca²⁺ concentration near the opening of the L-type Ca²⁺ channels, whereas SK channels responded to submembrane Ca²⁺concentrations. B, Expanded time scale of the AP and the associated Ca²⁺ transients identified inA by the asterisks. C, Isolated ionic currents underlying the generation of the single spike AP (left) and plateau-burst AP (right) identified in B.

To simulate the role of BK channels in generating plateau-bursting activity, we introduced these channels into the gonadotroph model cell. On the basis of our experimental results obtained from rat somatotrophs, the kinetics of BK channel activation and deactivation were dependent on the amplitude and kinetics of domain [Ca²⁺]_i (for details, see Materials and Methods). Introduction of BK channels into the model shifted the mode of AP firing from single spiking to plateau bursting (Fig. 11A, right panel), which was similar to that observed in spontaneously active somatotrophs. Specifically, the profile of the AP waveform was characterized by a sustained plateau potential during which multiple small-amplitude single spikes were observed (Fig.11B, right panel). In addition, the plateau-bursting activity in the gonadotroph model cell had a higher capacity to drive extracellular Ca²⁺ entry than the single-spiking activity observed in the absence of BK channels. Separation of the underlying currents indicated that rapid BK channel activation by domain [Ca²⁺]_i truncated the peak spike amplitude (Fig. 11C). This reduced the magnitude of DR channel activation, resulting in a decrease in the rate and magnitude of spike repolarization. Moreover, both experimental and mathematical simulations indicate that rapid deactivation of the BK channels by the clearance of domain Ca²⁺prevents them from fully repolarizing the membrane, allowing for the generation of the plateau potential (Fig. 10B,C).

DISCUSSION

In many neuronal, endocrine, and muscle cells, BK channels reduce membrane excitability and the associated Ca²⁺ signals by allowing the passive flux of K⁺ from the cell (Kaczorowski et al., 1996; Sah, 1996; Vergara et al., 1998). During AP firing the BK channels are gated by membrane depolarization, and this gating is facilitated by the concomitant increase in Ca²⁺ influx through VGCCs (Stefani et al., 1997). Once open, BK channels act in conjunction with other K⁺ channels to facilitate membrane hyperpolarization. This limits the duration of the AP waveform, reducing its capacity to drive extracellular Ca²⁺ entry (Lancaster and Adams, 1986;Lang and Ritchie, 1987; Strom, 1987). BK channel activation also generates the “fast” afterhyperpolarization observed during AP firing in some cell types (Lancaster and Adams, 1986). Accordingly, in many different cell types BK channel inhibition by pharmacological blockers (Strom, 1987; Lang and Ritchie, 1989; Pedarzani et al., 2000) or frequency-dependent inactivation (Shao et al., 1999) increases the duration of the AP waveform and the associated Ca²⁺ signals. Thus, by controlling the rate of membrane repolarization and the frequency of firing, BK channels act as a negative feedback pathway to limit AP-driven Ca²⁺ entry.

In rat somatotrophs, however, BK channel activation prolongs AP duration, which facilitates extracellular Ca²⁺ entry. This is supported by four lines of evidence. First, the introduction of rapid Ca²⁺ buffers into the cytosol of somatotrophs prevented BK channel activation and changed the profile of the AP waveform from plateau bursting to single spiking. Second, the inhibition of BK channels by the highly specific BK channel blockers, IBTX or paxilline, also shifted the profile of the AP waveform from plateau bursting to single spiking, which reduced its capacity to drive extracellular Ca²⁺ entry. Third, unlike somatotrophs, gonadotrophs express few BK channels and fire single spikes that are not altered by BAPTA, IBTX, or paxilline and have a low capacity to drive Ca²⁺ entry. Finally, the introduction of BK channels into a mechanistic biophysical model of rat gonadotrophs shifted the pattern of AP firing from single spiking to plateau bursting, which facilitated AP-driven Ca²⁺ entry.

On the basis of our experimental data, we propose the following mechanism for the generation of plateau bursting by BK channels in rat somatotrophs. Activation of VGCCs initiates the AP and generates the first spike. The ensuing Ca²⁺ entry through the VGCC rapidly activates a fraction of the BK channels that are closely associated with the VGCCs. This truncates the spike amplitude well below 0 mV and thereby limits the magnitude of DR channel activation. Together, the BK and DR channels partially repolarize the membrane, leading to the rapid deactivation of BK channels caused by the reduction in Ca²⁺influx through VGCCs. This leads to a reduction inI_BK, and, because the magnitude of DR channel activation is insufficient to repolarize the membrane fully, a plateau potential is generated. The plateau potential is attributable to the balance between the outward I_DRand I_BK and the inward current through VGCCs. The generation of the plateau potential results in the sustained influx of extracellular Ca²⁺ through VGCCs, resulting in the high-amplitude [Ca²⁺]_i transients in somatotrophs. The eventual repolarization of the plateau potential may be attributable to the recruitment of additional BK channels and/or activation of SK channels or Ca²⁺-activated Cl⁻ channels, both of which have been identified in pituitary cells (Ritchie, 1987; Korn et al., 1991), by the slower and lower amplitude rise in bulk [Ca²⁺]_i. Alternatively, Ca²⁺-dependent inhibition of VGCCs or the putative cyclic nucleotide-gated channels observed in rat somatotrophs (Tomic et al., 1999) may repolarize the plateau potential.

Critical to the generation of plateau bursting by BK channels is that they activate rapidly enough to truncate the peak spike amplitude, which in turn leads to a reduction in the magnitude of DR channel activation. In addition to responding to domain [Ca²⁺]_i, this indicates that the intrinsic voltage-dependent gating of the BK channels in somatotrophs also must be rapid. Consistent with this, in the absence of accessory β-subunits, which can slow channel activation (Dworetzky et al., 1996), the pore-forming α-subunit of the BK channel has been demonstrated to activate rapidly (<3 msec;Safronov and Vogel, 1998; Araque and Buno, 1999). In addition, in spinal neurons from Xenopus embryos the peak magnitude ofI_BK lagged behind the voltage-gated Ca²⁺ current by <50 μsec (Yazejian et al., 2000), indicating that BK channels can respond rapidly to domain [Ca²⁺]_i. Although the BKβ₂- and BKβ₃-subunits have been identified in the pituitary (Behrens et al., 2000), functional analysis indicates that the BKα-subunit is not coupled to any of the known BKβ-subunitsin vivo (Shipston et al., 1999). In addition, because association of the BKβ₂- and BKβ₃-subunits with the BKα-subunit is known to cause inactivation (Uebele et al., 2000; Xia et al., 2000), the lack of BK channel inactivation in rat somatotrophs further indicates that the α-subunit in these cells is not regulated by these β-subunits. It is, therefore, interesting to note that slowing BK channel activation by the association with BKβ-subunits, insertion into the membrane further from the VGCCs, or reducing their sensitivity to Ca²⁺ would prevent them from truncating the spike amplitude and reducing DR channel activation. This, in addition to their delayed activation, would allow them to facilitate membrane repolarization and limit AP-driven Ca²⁺ entry.

Our results indicate that rapid activation of BK channels by domain Ca²⁺ truncates spike amplitude. This limits DR channel activation, allowing for the generation of the prolonged plateau potential that characterizes the low-frequency phasic firing in rat somatotrophs. The ionic mechanisms underlying the generation of the rhythmic membrane depolarizations that generate the repetitive Na⁺–K⁺-dependent spiking during phasic firing also have been studied in other cell types. In Purkinje neurons (Raman and Bean, 1997) and “chattering” neocortical neurons (Brumberg et al., 2000) high-frequency burst firing (interburst frequency, 10–40 Hz) has been observed. This high-frequency burst firing is characterized by brief (<100 msec) rhythmic membrane depolarizations on which several high-amplitude fast spikes are observed (intraburst frequency, ∼300 Hz). The rhythmic depolarizations are attributable to the repetitive firing of fast afterdepolarizations that are generated by the persistent or resurgent activation of a tetrodotoxin-sensitive Na⁺current (Raman and Bean, 1997; Brumberg et al., 2000). A similar mechanism does not appear to operate in rat somatotrophs, because tetrodotoxin application has no effect on AP-driven [Ca²⁺]_i transients (Tomic et al., 1999) or plateau-bursting activity (F. Van Goor, unpublished observation).

In thalamocortical neurons, interactions between T-type VGCC and hyperpolarization-activated cationic channels generate low-frequency burst firing (interburst frequency between 0.5 and 10 Hz; Jahnsen and Llinás, 1984a,b; McCormick and Pape, 1990). This low-frequency burst firing is characterized by the rebound activation of T-type VGCC, which generates a transient Ca²⁺ spike with a duration of 100–200 msec. Superimposed on the T-type VGCC spikes are several fast Na⁺–K⁺-dependent spikes. Repolarization of the burst is attributable to the inactivation of the T-type VGCC channels and is followed by a hyperpolarizing overshoot because of a reduction in the depolarizing influence of hyperpolarization-activated cationic channels. In such cells, sustained membrane depolarizations inactivate T-type VGCC, which abolishes the underlying Ca²⁺ spikes and shifts the pattern of AP firing from bursting to single spiking (McCormick and Pape, 1990). In rat somatotrophs the activation of T-type VGCC appears to be important in initiating the first spike during plateau-bursting activity (Tomic et al., 1999; Van Goor, unpublished observations). In addition to acting as a pacemaker current, a noninactivated fraction of these channels may contribute to the plateau potential and participate in voltage-gated Ca²⁺ entry. However, further studies are required to identify the voltage-gated Ca²⁺ channel subtypes and their contribution to plateau bursting in somatotrophs. The role of VGCC inactivation during burst termination also requires further study.

In conclusion, these results indicate that the differential expression of BK channels between somatotrophs and gonadotrophs underlies the differences in the profile of the AP waveforms and the associated [Ca²⁺]_i signals. In rat somatotrophs, plateau-bursting activity generates high-amplitude [Ca²⁺]_i signals, whereas the single-spiking activity in gonadotrophs has a low capacity to drive Ca²⁺ entry. These differences in the pattern of AP firing and the associated [Ca²⁺]_i likely account for the differences in basal growth hormone and luteinizing hormone secretion in the absence of neuroendocrine regulation.