The ATP-sensitive potassium channel (K,,, channel) determines the membrane potential of pancreatic p-cells and plays a critical role in the regulation of insulin ...
0013.7227/95/$03.00/O Endocrinology Copyright 0 1995 by The Endocrme
Vol. Printed
136, No. 7 III U.S.A.
Society
Modulation of Adenosine Triphosphate-Sensitive Potassium Channel and Voltage-Dependent Channel by Activin A in HIT-T15 Cells* HIDE0 MOGAMI, MAKOTO MEGUMI FURUKAWA, AND Department of Cell Biology, Maebashi 371, Japan
KANZAKI, ROM1 ITARU KOJIMA
Institute
for Molecular
NOBUSAWA,
and Cellular
YOU-QING
Regulation,
Gunma
Calcium
ZHANG, Uniuersity,
ABSTRACT The ATP-sensitive potassium channel (K,,, channel) determines the membrane potential of pancreatic p-cells and plays a critical role in the regulation of insulin secretion. The present study was conducted to investigate the effect of activin A, a member of the transforming growth factor-p supergene family, on the KATP channel in HIT-T15 clonal hamster insulinoma cells. In an excised inside-out patch, ATP-sensitive currents with a single channel conductance of approximately 20 picosiemens were observed. In an outside-out patch, currents with identical unitary conductance were also observed. In either case, the currents were augmented by diazoxide and blocked by glibenclamide, verifying that they were KATp channel currents. When K ATP channel currents were monitored in an outside-out patch, activin A added to the bath solution inhibited KATP channel currents. currents were restored, Upon removal of activin A, the KATP channel
suggesting that the inhibition was not due simply to spontaneous disappearance of channel activity (run-down). The KATP channel activity was markedly reduced after the addition of activin A and was reversed by diazoxide. Besides the inhibition of KATp channel, activin A increased, in a perforated patch, the amplitude of the inward Ba” current in response to a depolarizing pulse from -40 to +lO mV. Under the current clamp condition, activin A induced gradual depolarization, followed by a burst of action potentials. Activin-mediated action potentials were accompanied by an elevation of the cytoplasmic free calcium concentration. These results indicate that activin A causes depolarization of the plasma membrane by inhibiting the activity of the KATp channel. In addition, activin A directly modulates the voltage-dependent calcium channel and augments calcium entry. (Endocrinology 136: 2960-2966, 1995)
A
activin A induces hydrolysis of phosphoinositides in murine erythroleukemia cells (10) and isolated rat hepatocytes (11). In addition, activin A increasesthe cytoplasmic free calcium concentration ([Ca’+],) in pituitary tumor cells by stimulating calcium entry (12). These effects are unique among the ligands belonging to the TGFP supergene family (8). In rat pancreatic islets, activin A is capable of stimulating insulin secretion (13). The effect of activin A is unique in several respects. First, activin A induces a biphasic secretory response of insulin in the presence of a subthreshold concentration of ambient glucose (14). Also, a low concentration of activin A, which does not augment insulin secretion by itself, markedly potentiates glucose-induced insulin secretion (14). Second, activin A increases[Ca2+], in islet cells (15), which is blocked by either removal of extracellular calcium or addition of nifedipine, an inhibitor of voltage-dependent calcium channel (16). Additionally, the activin-induced elevation of [Ca2’], was blocked by diazoxide (15), an opener of the ATP-sensitive potassium channel (K,,,, channel) (17). Thesetwo observations suggestthat activin A, either directly or indirectly, activates the voltage-dependent calcium channel. Recently, Verspohl et al. (18) reported that activin A reduced [s6Rb] efflux in islets. This observation raises the interesting possibility that activin A causesdepolarization of the plasma membrane by inhibiting the activity of a certain classof potassium channels. In pancreatic p-cells, K,,,, channel plays a pivotal role in the regulation of insulin secretion. The resting membrane potential is determined mainly by the activity of the K,,,, channel (see Ref. 19 for review) and
A is a homodimeric protein structurally related to transforming growth factor-p (TGFP) (1). This protein is considered to be a member of the TGFP supergene family and elicits diverse effects in various cell systems (2). The receptor for activin A was determined by molecular cloning (3), and the primary structure suggested the existence of serine-threonine kinase activity located in its intracellular domain. This receptor, designated the type II activin receptor (3), was purified to homogeneity and was shown to possessserine-threonine and tyrosine kinase activities (4). As in the caseof the TGFP receptor system (for review, seeRef. 5), the activin receptor system was more complex than previously thought, and there existed another type of serinethreonine receptor kinase, designated the type I receptor (6, 7). It is postulated that dimers of type I and type II receptors are necessary for activin A to generate intracellular signals (for review, see Ref. 8). In some cell systems, the type III activin receptor is characterized (9). Despite the fact that receptor molecules have been characterized, only limited information is available at present concerning the postreceptor signaling system activated by activin A. We have shown previously that activin A activates the calcium messenger system in at least some types of target cells. Thus, CTIVIN
Received January 31, 1995. Address all correspondence and requests for reprints to: Itaru Kojima, M.D., Institute for Molecular and Cellular Regulation, Gunma University, Maebashi 371, Japan. * This work was supported by a Grant-in-Aid for Scientific Research from the Ministry of Education, Science, and Culture of Japan.
2960
ACTIVIN
ACTION
agents that inhibit the activity of the K,,, channel augment insulin secretion. For example, glucose increases insulin secretion mainly by inhibiting the activity of the K,,, channel (19). Also, antidiabetic agents, sulfonylureas, stimulates insulin secretion by blocking the K,,, channel. In the present study, we studied the ionic mechanism by which activin A stimulates calcium entry in a hamster insulinoma cell line, HIT-T15 cells. As there is a possibility that activin A may modulate the potassium channel (18), and the resting membrane potential of the p-cell is determined mainly by the activity of the K,,, channel (19), our study was focussed on the effect of activin A on the K,,,, channel. The results indicate that activin A inhibits the activity of the K +,ri, channel and thereby causes depolarization. Activin A also enhances the activity of the L-type voltage-dependent calcium channel. Hence, activin A acts on the ion channels modulated by glucose, a primary regulator of insulin secretion. Materials
and
Methods
Recombinant human activin A was provided by Dr. Research Laboratory, Ajinomoto (Kawasaki, Japan). purchased from Dojin Chemicals (Kumamoto, Japan). tained from DuPont-New England Nuclear (Boston, glibenclamide, and mannoheptulose were purchased Chemical Co. (St. Louis, MO).
Y. Eto of Central Fura-Z/AM was Na[‘Z”IJ was obMA). Diazoxide, from Sigma
Cell culture HIT-T15 cells (201, provided by Prof. T. Takeuchi of the Institute for Molecular and Cellular Regulation, Gunma University, were cultured in Dulbecco’s Modified Eagle’s Medium containing 2 mM glutamine and 10% fetal calf serum in a humidified atmosphere of 5% CO, and 95% air at 37 C. For measurement of [‘?]activin A binding, cells were cultured in a 24-well plate (Falcon Lincoln Park, NJ). For the patch clamp recording, cells were seeded onto a glass coverslip.
of insulin
were incubated for 60 min in Krebs1.25 mM calcium, 2.7 mM glucose, washed and then incubated for 2 h A or a high concentration of glucose, period was measured by RIA, as
of [‘Z5ZIactivin
A binding
Activin A was iodinated using lactoperoxidase, as described previously (21). The specific activity of [‘2”IJactivin A ranged from 50--100 pCi/mg. For the measurement of [ “51Jactivin A binding, cells were incubated for 3.5 h at room temperature with [“‘Ilactivin A in the presence and absence of an excess of unlabeled activin A. Cells were washed twice with PBS, then lysed by adding 0.5 M NaOH, and assoTABLE
1. Types
of membrane
patches
Cell-attached Inside-out
mode mode
Outside-out
mode whole
and data analysis
We used the single channel variation of the whole cell patch clamp channel events technique (23) for the analysis of K AT,, channel unitary and measured changes in membrane currents and membrane potential using the perforated mode of the patch clamp technique (24). The membrane patches used in this study are summarized in Table 1. All data were obtained using a computer-based amplifier system (EN-9) controlled by E9 screen software (HEKA, Lambrecht, Germany) on an Atari (Sunnyvale, CA) computer. Single channel recording was begun 1 min later after establishment of the single channel variation. Current clamp recording was started after the series resistance fell below 50 mR. Voltage clamp recordings, which were compensated by capacity and leak current subtraction, were started after the series resistance fell below 20 mfl considering the voltage error. All experiments were carried out at 26- -30 C. Analysis of single channel recording was performed by using TAC program (HEKA). The total number of functional channels (N) in the patch was estimated by observing the number of peaks detected on the amplitude histogram. As an index of channel activity, NPo (number of channels times the open probability) is calculated as
used in this
where T is the total record time, n is the number of channels open, and t, is the recording time during which n channels are open. Therefore, NPo can be calculated without making assumptions about the total number of channels in a patch or the open probability of a single channel. For simultaneous measurement of [Caz+Jc, cells were loaded with furaby incubation with 2 FM fura-2/AM for 30 min at room temperature. Cells were then washed twice with standard extracellular solution, as described below, and fluorescence from a single cell was monitored using CAM-230 (Nihon Bunko, Tokyo, Japan). The ratio of emissions obtained by excitation wave lengths of 340 and 380 nm (340/380 ratio) was used as an index of [Ca”],. The standard extracellular solution contained 137 mM NaCl, 5 mM KCl, 1 mM MgCI,, 2 mM CaCl,, 5 mM NaHCO,, 2.7 mM glucose, and 10
study Advantage
TYP=
Perforated
measurement
secretion
Cells cultured in a 24.well plate Ringer bicarbonate buffer containing and 0.1% BSA (Sigma). Cells were in medium containing either activin and insulin released during the described previously (14).
Measurement
ciated radioactivity was counted (21). Affinity cross-linking of the [“sllactivin A-binding protein was performed using disuccinimidyl suberate (DSS). Confluent cells grown in a IO-cm dish were washed three times with the binding buffer (Dulbecco’s Modified Eagle’s Medium containing 0.2% BSA and 0.01% NaNal and incubated in 5 ml I’2”IJactivin A solution (10 rig/ml in the same buffer) in the presence and absence of unlabeled activin A at 4 C for 3.5 h. After washing once with the binding buffer and twice with ice-cold PBS, DSS was added to a final concentration of 0.5 mM to cross-link the bound [‘Z”IJactivin A, and cells were then incubated at 4 C for 30 min. The reaction was quenched with 1 ml 20 mM Tris-HCl (pH 7.4) containing 0.15 M NaCl and 1 mM EDTA. After 5 min, the cells were harvested and solubilized in 20 mM Tris-HCl (pH 7.4) containing 0.1% Triton X-100, 1 ml phenylmethylsulfonylfluoride, 1 mM EDTA, 5 mM iodoacetamide, 10 mg/ml pepstatin A, and 1000 U/ml trypsin inhibitor. The cell suspensions were stirred gently at 4 C for 30 min and centrifuged at 12,000 rpm at 4 C for 10 min. The supernatants were mixed with an equal volume of 2-fold concentrated Laemmli electrophoresis sample buffer (221 in the presence of 50 mM mercaptoethanol. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) of the affinity-labeled samples was carried out using 7.5% gels. After SDS-PAGE, the gels were dried and exposed to Kodak XAR film (Eastman Kodak, Rochester, NY) at -80 C for 5 days.
Electrophysiological
Materials
Determination
IN HIT CELLS
cell mode
Single channel current can be measured in an intact cell Effect of the agent added to the cytoplasic side of the plasma membrane can be determined Effect of the agent added to the extracellular side of the plasma membrane can be measured Whole cell current can be measured without losing intracellular molecules
2962
ACTIVIN
ACTION
mM HEPES-NaOH (pH 7.4). The solution for measurement of the voltage-gated calcium channel contained 115 mM NaCl, 5 mM CsCl, 1 mM M&l,, 10 mM BaCl,, 10 mM tetraethylammonium chloride, 2.7 mM elucose. and 10 mM HEPES-NaOH (uH 7.4). The standard uatch elecGode solution contained 95 mM pota&ium aspartate, 40 mM’KC1,5 mM NaCl, 1 mM M&l,, and 10 mM HEPES-N-methyl-o-glucamine (pH 7.2). For blocking the outward currents, K ions were replaced with Cs ions. For the outside-out mode, 3 mM EGTA, 0.2 mM ATP, and 0.05 mM GTP were added.
A binding
in HIT-T15
l
1995 No 7
100
cells
In the first set of experiments, we examined whether activin A affected insulin secretion in HIT-T15 cells. Insulin secretion rates in the presence of 2.7 and 25 mM glucose were 34.5 ? 4.0 and 64.1 2 5.1 pU/well.2 h, respectively (mean + SE; n = 4). When cells were incubated with 2 nM activin A in the presence of 2.7 mM glucose, the secretion rate was 69.4 -t 5.7 pU/well.2 h (n = 4). We then characterized functional activin receptors in HIT-T15 cells by measuring [i2”I]activin A binding. [‘2511Activin A was bound to HIT-T15 cells specifically. The binding was displaced by unlabeled activin A in a dose-dependent manner (Fig. 1A). Figure 1B depicts the Scatchard plot of the [i2”I]activin A binding onto HIT-T15 cells. The apparent K, value of the binding was 2.4 X 10-i’ M, and the total number of binding sites was 6100/tell. Figure 2 depicts affinity cross-linking of [‘*“I]activin A binding in HIT-T15 cells. Two protein bands with molecular masses of 65 and 85 were specifically labeled with [‘251]activin A. Based on their mol wt, they might correspond to the type I and type II activin receptors, respectively. Effect of activin
Endo Vol 136.
50
Results [‘251]Activin
IN HIT CELLS
A on KATp channel
In p-cells, the resting membrane potential was determined largely by the activity of the KATF channel (19). We, therefore, examined the effect of activin A on the K,,, channel. To characterize K,,, channel activity, we performed single channel current recording obtained in both excised insideout and outside-out membrane patches. The solutions at the extracellular and intracellular sides of the membrane contained 5 and 135 mM potassium, respectively, and the membrane potential was held at 0 mV. In an inside-out patch, the addition of 3 mM ATP resulted in complete inhibition of outward single channel currents, and recovery of the currents occurred after the washout of ATP (Fig. 3A). The outward currents in an excised inside-out membrane patch were enhanced by 400 PM diazoxide, an opener of the K,,, channel (17), and were abolished by 1 PM glibenclamide, an inhibitor of the K,,, channel (19) (data not shown). Figure 3B shows the single channel current-voltage (I-V) relationship under the conditions described above. The slope conductance was 21 picosiemens, which is in agreement with that for the K ATF channel (19). In an outside-out patch, single channel currents were also observed (Fig. 3C). In this condition, calcium was chelated, and membrane potential was fixed at 0 mM. Therefore, the voltage-dependent and calcium-activated potassium channel was blocked. It was likely that among the various potassium channels, only the K,,, channel was operative. Consistent with this notion, single
I,
0
10-l’
lo-lo
10”
loa
[Ligand] (M)
0.03
0.02
0.01
5
10
[125 I]Activin A Bound (PM) FIG. 1. Bindingof[ i2sI]activin A in HIT-15 cells. [‘“‘I]Activin A binding was measured in intact HIT-T15 cells, as described in Materials and Methods. A, Displacement of I ‘251]activin binding by unlabeled activin A. B, Scatchard plot of [‘2sIlactivin A binding.
channel currents detected in the outside-out patch were activated by diazoxide and blocked by glibenclamide (Fig. 30. Figure 3D shows the I/V relationship for single channel current obtained in an outside-out patch, which was nearly identical to that shown in Fig. 38. The slope conductance was approximately 20 picosiemens. Hence, currents detected in both patch configurations were identical. To study the effect of activin A on K,,, channel, we measured K,,, currents in an excised outside-out configuration. The pipette solution contained 0.2 mM ATP and 0.05 mM GTP, and the holding potential was 0 mV. As shown in Fig. 4A, addition of 2 nM activin A resulted in inhibition of opening of the K,,, channel. Upon removal of activin A, opening of the K,,, channel was restored gradually, suggesting that the inhibition was
ACTIVIN
ACTION
IN HIT CELLS
2963
97 69 46
-
unlabeled activin A
-
+
FIG. 2. Affinity cross-linking of [1251]activin A binding sites. [‘2511Activin A was cross-linked using DSS in the presence and absence of unlabeled activin A. Proteins were separated by SDS-PAGE under reducing conditions.
not due simply to run-down. Similarly, activin A-mediated inhibition of the KATp channel was immediately reversed by the subsequent addition of diazoxide, and the addition of gli-benclamide blocked opening of the KArr channel (Fig. 4B). Figure 4C demonstrates the changes in the channel activity (NPo) of the K,, channel after the addition of activin A. NPo was markedly reduced after the addition of activin A, whereas NPo did not change in the absenceof activin A. The effect of activin A was observed in the absenceof GTP in the pipette solution, but was not detected in the absence of ATP in the pipette solution (data not shown). When K,, channel currents were measured in a cell-attached patch, activin A added outside the patch was ineffective in inhibiting K,, channel currents (data not shown). Effect of activin
A on voltage-dependent
calcium
channel
To examine the possibility that activin A modulated the voltage-dependent calcium channel, inward Ba2+ current was measured in a perforated mode of patch clamp. Outward currents were blocked. As shown in Fig. 5A, when membrane potential was increased from a holding potential of -40 to 10 mV, an inward Ba2+current was observed. Note that the T-type voltage-dependent calcium channel did not operate under this condition. In agreement with this notion, the inward Ba2+ current was blocked by 1 PM nifedipine (data not shown). Treatment of the cells with activin A for 1 min enhanced the amplitude of the inward Ba2+current 45%. Figure 5B depicts the I-V relationship of the inward Ba2+ current in naive and activin A-treated cells. Activin A enhanced inward Ba2+ current without shifting the I-V relationship. Effect of activin
A on membrane
potential
To determine the effect of activin A on cellular calcium, we monitored changesin membrane potential in HIT cells by the
I
/ -80
-40
0
40
mV
FIG. 3. Identification of KATp channel current in inside-out and outside-out patches. Currents were measured in inside-out (A and B) and outside-out (C! and D) patches, as described in Materials and Methods. Membrane potential was held at 0 mV. In C, 400 @ d&oxide and 1 /.LM glibenclamide were added, as indicated. Current-voltage relationships for currents obtained in inside-out (B) and outside-out (D) patches were shown. The inside-out patch depicts outward currents as downward deflections. The horizontal bar marks the zero count level W, closed state). Recordings were flltered at 200 Hz low pass. The outside-out patch depicts outward currents as upward deflections.
2964
ACTIVIN
ACTION
IN HIT CELLS
Endo. Vol 136.
1995 No 7
A
mV
1
T L 0 P
PA FIG. 5. Effect of activin A on the voltage-dependent calcium channel. Voltage-dependent calcium channel current was measured in a perforated patch using Ba2+ as a charge carrier. A, Membrane potential was changed from a holding potential of ~40 to 10 mV in cells before and 1 min after the addition of 2 nM activin A. B, Voltage-dependent calcium current was measured by various voltage jumps, and the IN relationship was measured in cells before (0) and 1 min after the addition of 2 nM activin A (0).
50
Before
Afler
FIG. 4. Effect of activin A on KATp channel. Transmembrane potassium currents were monitored in the outside-out patch. Activin A (2 nM) was added to the bath, as indicated (A). B, Diazoxide (400 wM) and glibenclamide (1 pM) were added sequentially after the addition of activin A. C, Cells were treated with 2 nM activin A (H) or saline (01, current was measured as described in A. NPo was and KATp channel calculated before and 3 min after the addition of 2 nM activin A. Statistical significance was evaluated by Student’s t test. Values are the mean 2 SE for eight experiments. *, P < 0.01 vs. before treatment.
current clamp technique in a perforated mode of patch clamp. The resting membrane potential was -60 to -65 mV in the presence of 2.7 mM glucose. The addition of 2 nM activin A resulted in a gradual increase in membrane potential, followed by a burst of action potentials (Fig. 6A). In most of the cells, the burst of action potentials continued persistently even though the frequency of action potentials fluctuated. In some small number of cells (-5%), oscillation of membrane potential was observed. Thus, the burst of action potential was terminated spontaneously within a few minutes, and membrane potential returned to the basal level, followed by periodical depolarization and the appearance of a burst of action potentials (Fig. 6B). Note that activin-induced depolarization was observed in the presence of 10 mM
mannoheptulose (data not shown). Figure 7 demonstrates the result of simultaneous recording of changes in membrane potential and [Ca”],. Activin A-induced action potentials were accompanied by an elevation of [Ca2+lc. Removal of extracellular calcium or addition of cobalt abolished action potentials (data not shown).
Discussion The present study was conducted to elucidate the mechanism by which activin A, a member of the TGFP supergene family, increases [Ca2+lc in a clonal insulinoma cell line, HIT-T15 cells. HIT-T15 cells express specific binding sites for activin A. [‘251]Activin A binding is replaced by unlabeled activin A, and Scatchard analysis reveals the existence of high affinity binding sites. As activin A exerts its effect on cellular calcium metabolism at concentrations of about 1O-9 M, it seems reasonable that activin A exerts its action via its own receptor system. As in other cells (8), affinity crosslinking demonstrates the existence of type I and type II receptors in HIT cells. Consistent with our previous observation that activin A
ACTMN
A
ACTION
Activin
I
1 min
B
Activin
FIG. 6. Changes in membrane potential in response to activin A. Membrane potential was monitored under a current clamp condition. Activin A (2 nM) was added as indicated by the arrow (A and B). In some cells, membrane potential was fluctuated (B).
I
Aelivin
A
- 60
1 min
FIG. 7. Effects of activin A on membrane potential and cytoplasmic free calcium. Membrane potential was monitored in a fura-2-loaded cell, and changes in membrane potential and [Ca”], were monitored simultaneously. Activin A (2 nM) was added as indicated by the arrozu.
evokes elevation of [Ca2’], in rat islet cells, activin A elevates [Ca2’], in HIT-T15 cells, which is dependent on calcium entry. The current clamp method provides evidence that activin A increases calcium entry via the voltage-dependent calcium channel; activin A causes depolarization of the plasma membrane and induces a burst of action potentials,
IN HIT
CELLS
2965
which is blocked by cobalt and nifedipine. Of particular importance is the fact that in an outside-out configuration of patch clamp recording, activin A reduces K,,, channel currents, which are sensitive to diazoxide and glibenclamide. Both ATP-sensitive and Ca2+- and voltage-activated potassium channels exist in insulin-secreting cells (25). In our experimental conditions, however, membrane potential is fixed at 0 mV, and moreover, [Ca2’] in the pipette solution is chelated by EGTA. Therefore, the Ca2+- and voltage-activated potassium channel is not o erational(25,26). Also, as mentioned by Dunne et al. (261, Ca pi - and voltage-dependent potassium current can be easily discriminated from K,,, channel current by the conductance. In fact, the unitary conductance of single channel current is identical to that of the K ATP channel detected in an inside-out patch. Hence, it is clear that activin A inhibits the activity of the KATP channel. It should be emphasized that the effect of activin A is observed in the outside-out mode, but not in the cell-attached patch when activin A is added to the bath solution. This raises the possibility that activin A inactivates the K,,, channel by a direct and presumably receptor-operated mechanism, rather than by involving a second messenger. In p-cells, glucose inactivates the KATP channel by altering the cellular level of ATP or the ATP/ADP ratio (19,27). Likewise, glucagon-like peptide-l inhibits the K,,, channel by a mechanism involving CAMP as a messenger (28). In contrast, activin A does not elevate intracellular CAMP (15). The present report is the first demonstration that a ligand inhibits the activity of the KATP channel by a direct, presumably receptoroperated, mechanism. With regard to the receptor-operated mechanism, it has been reported that the K,,, channel is activated by galanine (26) and somatostatin (29) through a mechanism involving a pertussis toxin-sensitive G protein. Although we measured the effect of activin A in the presence of GTP in the pipette solution, GTP was not absolutely necessary. Therefore, activin A may inactivate the K,,, channel by a mechanism independent of G protein. At present, the precise mechanism for activin receptor-mediated inhibition of the K,,, channel is not known. The activin A receptor system is unique, in that both type I and type II receptors possess predicted serine-threonine kinase. The activin receptor system in HIT-15 cells is comprised of type I and type II receptors. As removal of ATP from the pipette solution attenuates the effect of activin A on the KAT1, channel, it is likely that receptor kinase activity is required for the operation of the activin receptor system. Takano et al. (30) reported recently that activin A modulates the activity of ion channels and augments calcium entry in human pituitary tumor cells. Specifically, activin A depolarizes the cell by stimulating Nat entry via the nonselective cation channel and increases the amplitude of L-type voltage-dependent calcium channel. Although the ionic control is slightly different between p-cells and pituitary cells, activin A modulates directly and indirectly various types of ion channels in the plasma membrane. Further study is necessary to elucidate the mechanism by which activin receptors regulate the K,,, channel. In addition to the effect on the K,,, channel, activin A acts on L-type voltage-dependent calcium channel and increases the Ba2+ current. Hence, activin A opens the voltage-dependent calcium channel by depolarizing the plasma membrane
ACTMN
2966
ACTION
and augments calcium entry by directly modulating the Ltype voltage-dependent calcium channel. Taken together, activin A is a unique agonist of insulin secretion, and its action resembles that of glucose (25, 31). Activin A is an agonist that mimics the action of fuel stimulator glucose on p-cell calcium metabolism. In this regard, we recently observed that activin A is capable of stimulating insulin secretion from islets in the absence of ambient glucose (Furukawa, M., H. Mogami, R. Nobusawa, and I. Kojima, submitted for publication). In addition, glucagon-like peptide-1, which cannot induce insulin secretion in the absence of glucose, stimulates insulin secretion in glucose-free medium when activin A is present. Such effects of activin A result from its ability to cause depolarization of the plasma membrane by inhibiting the K,,, channel independent of glucose metabolism. Our unpublished observation indicated that activin A also induces a burst of action potentials in isolated rat p-cells. Although the present results were obtained in an insulinoma cell line, HIT T-15 cell, it is likely that activin A modulates ion channels in normal p-cells in a similar manner.
IN HIT CELLS
11. 12.
13.
14.
15.
16. 17.
18. 19.
Acknowledgment The authors are grateful of the manuscript.
to Dr. Norio
Kawamura
for critical
reading
References 1. Vale W, Rivier J, Vaughan J, McClintock R, Corrigan A, Woo W, Karr D, Spiess J 1986 Purification and characterization of an FSH releasing protein from porcine ovarian follicular fluid. Nature 321: 776 -779 2. Depaolo LV, Bicsak TA, Erickson GF, Shimasaki S, Ling N 1991 Follistatin and activin A: a potential intrinsic regulatory system within diverse tissues. Proc Sot Exp Biol Med 198:500-512 3. Mathews L, Vale W 1991 Expression cloning of an activin receptor, a predicted transmembrane serine kinase. Cell 65:973-982 4. Nakamura T, Sugino K, Kurosawa N, Sawai M, Takio K, Eto Y, Iwashita S, Muramatsu M, Titani K, Sugino H 1992 Isolation and characterization of activin receptor from mouse embryonal carcinoma cells. J Biol Chem 26718924-18928 5. Massague J, Attisano L, Wranna JL 1993 The TGF-/3 family and its composite receptor. Trends Cell Biol 4:172-178 6. Attisano L, Carcamo J, Ventura F, Weis FMB, Massague J, Wranna J L 1993 Identification of human activin and TGF-/3 type I receptors that form heteromeric kinase complexes with type II receptors. Cell 75:671-680 7. Franzen P, de Dijk P, Ichijo H, Yamashita H, Schulz P, Heldin C-H, Miyazono K 1993 Cloning of a TGF-p receptor that form a heteromerit comulex with the TGF-B tvue II receutor. Cell 75:681-692 8. Mathews i 1994 Activin re&pt& and ce?lular signaling by the receptor serine kinase family. Endocr Rev 15:310-325 9. McCarthy S, Bicknell R 1994 Activin A binds to a heterotrimericreceptor complex on the vascular endothelial cell surface. J Biol Chem 269:3909-3912 10. Shibata H, Ogata E, Eto Y, Shibai H, Kojima I 1987 Erythroid differentiation factor stimulates polyphosphoinositides breakdown
20.
21.
22. 23.
24.
25.
26.
27. 28.
29.
30. 31.
Endo. 1995 Vol 136 . No 7
in Friend leukemia cells. Biochem Biophys Res Commun 146: 187-193 Mine T, Kojima I, Ogata E 1989 Stimulation of glucose production by activin A in isolated hepatocytes. Endocrinology 125:586-591 Takano K, Yamashita N, Kojima I, Kitaoka M, Teramoto A, Ogata E 1992 Effect of activin A on intact FSH secretion and intracellular Ca2+ concentration in human FSH-secreting pituitary adenoma cells. Biochem Biophys Res Commun 182:1408-1415 Totsuka Y, Tabuchi M, Kojima I, Ogata E 1988 A novel action of activin A: stimulation of insulin secretion in rat pancreatic islets. Biochem Biophys Res Commun 156:335-339 Yasuda H, Inoue K, Shibata H, Takeuchi T, Eto Y, Hasegawa Y, Sekine N, Totsuka Y, Mine T, Ogata E, Kojima I 1993 Existence of activin A in A- and D-cells of rat pancreatic islets. Endocrinology 133:624-630 Shibata H, Yasuda H, Sekine N, Mine N, Totsuka Y, Kojima I 1993 Activin A increases cytoplasmic free calcium concentration in rat pancreatic islet. FEBS Lett 329:194-198 Hachisu M, Pappano P 1983 Dihydropyridine calcium antagonists. J Pharmacol Exp Ther 225:112-120 Trube G, Rorsman P, Ohno-Shosaku T 1986 Opposite effects of tolbutamide and diazoxide on the ATP-dependent K+-channel in pancreatic P-cells. Pflueger Arch 407~493-499 Verspohl EJ, Ammon HPT, Wahl MA 1993 Activin A: its effect onral pancreatic islets and the mechanism involved. Life Sci 53:1069-1078 Ashcroft FM, Rorsman P 1991 Electrophysiology of the pancreatic p-cell. Prog Biophys Mol Biol 54:87-143 Santerre RF, Cook RA, Crisel RMD, Sharp JD, Schmidt RJ, Williams DC, Wilson CP 1981 Insulin synthesis in a cloned cell line of simian virus 40-transformed hamster pancreatic beta cells. Proc Nat1 Acad Sci USA 78:4339-4343 Yasuda H, Mine T, Shibata H, Eto Y, Takeuchi T, Asano S, Kojima I 1993 Activin A; an autocrine inhibitor of initiation of DNA synthesis in rat hepatocytes. J Clin Invest 92:1491-1496 Laemmli UK 1970 Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227~680-685 Hamil OP, Marty OM, Neher E, Sakmann B, Sigworth FJ 1981 Improved patch-clamp techniques for high-resolution current recording from cells and cell-free membrane. Pflueger Arch 391: 85-99 Rae J, Cooper K, Gates P, Watsky M 1991 Low access resistance perforated patch recordings using amphotericin BJ. Neurosci Method 37~15-25 Findlay I, Dunne MJ, Petersen OH 1985 ATP-sensitive inward rectifier and voltageand calcium-activated K channels in cultured pancreatic islet cells. J Membr Biol 88:165-172 Dunne MJ, Bulett MJ, Li G, Wollheim CB, Petersen OH 1989 Galanin activates nucleotide-sensitive K channel in insulin-secreting cells via apertussis toxin-sensitive G protein. EMBO J 8:413-420 Cook D, Hales CN 1984 Intracellular ATP directly blocks K channels in pancreatic B-cells. Nature 311:271-273 Holz GG, Kuhtreiber WM, Habener JF 1993 Pancreatic beta-cells are rendered glucose-competent by the insulinotropic hormone glucagon-like peptide(7-37). Nature 361:362-365 Fosset M, Schmid-Antomarchi H, DeWeillie JR, Lazdunski M 1988 Somatostatin activates ATP-sensitive K channel by a G proteindependent mechanism. FEBS Lett 242:94-96 Takano K, Ogata E, Yamashita N 1994 Effects of acrivin A on ionic channels in human FSH-secreting tumor cells. J Physiol 474:65-73 Smith PA, Rorsman P, Aschcroft FM 1989 Modulation of dihydropyridine-sensitive Ca*+ channels by glucose metabolism in mouse pancreatic P-cells. Nature 342:550-553
