1. Top of page
  2. Abstract
  6. Acknowledgements
  • 1
    The perforated patch whole-cell configuration of the patch-clamp technique was applied to superficial glucagon-secreting α-cells in intact mouse pancreatic islets.
  • 2
    α-cells were distinguished from the β- and δ-cells by the presence of a large TTX-blockable Na+ current, a TEA-resistant transient K+ current sensitive to 4-AP (A-current) and the presence of two kinetically separable Ca2+ current components corresponding to low- (T-type) and high-threshold (L-type) Ca2+ channels.
  • 3
    The T-type Ca2+, Na+ and A-currents were subject to steady-state voltage-dependent inactivation, which was half-maximal at −45, −47 and −68 mV, respectively.
  • 4
    Pancreatic α-cells were equipped with tolbutamide-sensitive, ATP-regulated K+ (KATP) channels. Addition of tolbutamide (0·1 mm) evoked a brief period of electrical activity followed by a depolarisation to a plateau of −30 mV with no regenerative electrical activity.
  • 5
    Glucagon secretion in the absence of glucose was strongly inhibited by TTX, nifedipine and tolbutamide. When diazoxide was added in the presence of 10 mm glucose, concentrations up to 2 μm stimulated glucagon secretion to the same extent as removal of glucose.
  • 6
    We conclude that electrical activity and secretion in the α-cells is dependent on the generation of Na+-dependent action potentials. Glucagon secretion depends on low activity of KATP channels to keep the membrane potential sufficiently negative to prevent voltage-dependent inactivation of voltage-gated membrane currents. Glucose may inhibit glucagon release by depolarising the α-cell with resultant inactivation of the ion channels participating in action potential generation.

It is well established that pancreatic islets are electrically excitable and utilise electrical signals to couple variations of the blood glucose concentration to stimulation or inhibition of hormone release. Applying the patch-clamp technique, the ionic conductances participating in the generation of the β-cell electrical activity have been characterised both in dispersed islet cell preparations (Ashcroft & Rorsman, 1989) and in intact islets (Göpel et al. 1999a, b). These studies have revealed that electrical activity and insulin secretion follows the closure of ATP-regulated K+ channels (KATP channels) with resultant membrane depolarisation, activation of voltage-gated L-type Ca2+ channels and an associated increase in the cytoplasmic Ca2+ concentration ([Ca2+]i).

Experiments on glucagon-secreting α-cells in preparations of dispersed guinea-pig and rat α-cells maintained in tissue culture indicate that they are also electrically excitable and generate Na+- and Ca2+-dependent action potentials in the absence of glucose (Rorsman & Hellman, 1988; Gromada et al. 1997). Somewhat surprisingly, rat α-cells contain KATP channels at high density (Bokvist et al. 1999) but their role in glucagon secretion is enigmatic. Electrophysiological data from α-cells in mouse pancreatic islets (a traditional preparation for studies on islet electrophysiology) are scarce. However, recent experiments have revealed that they are also electrically active at low glucose concentrations (Yoshimoto et al. 1999). Measurements of [Ca2+]i in α-cells in intact mouse islets further suggest that elevation of the glucose concentrations leads to a reduction of [Ca2+]i (Berts et al. 1996; Nadal et al. 1999). Given these features, in particular the inverse glucose dependence, it seems likely that the electrophysiological characteristics of the α-cell must be rather different from those of the β-cell.

Here we have applied the patch-clamp technique to α-cells within intact mouse pancreatic islets (Göpel et al. 1999a). In this and the accompanying study (Göpel et al. 2000), we demonstrate that the α-cells can be distinguished from the β- and δ-cells by: (1) being electrically silent in the presence of insulin-releasing glucose concentrations; (2) containing a large TTX-sensitive Na+ current which inactivates at intermediate voltages; (3) possessing a TEA-resistant A-current; and (4) being equipped with two types (T- and L-type) of voltage-dependent Ca2+ channel. We further describe that the voltage-dependent Na+ channels are critical to glucagon secretion and that mouse α-cells, like their rat counterparts, contain KATP channels. Based on these data we propose a model for glucose regulation of glucagon release in which KATP channels and inactivation of voltage-dependent Na+ channels play an important role.


  1. Top of page
  2. Abstract
  6. Acknowledgements

Preparation of islets and electrophysiology

The methods are as described in detail in the accompanying paper (Göpel et al. 2000). Briefly, mice were stunned by a blow to the head and killed by cervical dislocation, and islets isolated by collagenase digestion (approved by the ethical committee of Lund University). Glucagon-secreting α-cells were functionally identified by the presence of both an A-current and a TTX-sensitive Na+ current. The perforated patch whole-cell technique was employed using electrodes containing 0·24 mg ml−1 amphotericin B and a pipette-filling solution composed of (mm): 76 K2SO4, 10 KCl, 10 NaCl and 5 Hepes (pH 7·35 with NaOH) for all experiments. The extracellular solution usually consisted of (mm): 140 NaCl, 3·6 KCl, 2 NaHCO3, 0·5 NaH2PO4, 0·5 MgSO4, 5 Hepes (pH 7·4 with NaOH), 2·6 CaCl2 and, unless otherwise indicated, 10 D-glucose. The Ca2+ currents were recorded using 2·6 mm Ba2+ as the charge carrier (Ca2+ being omitted) in the simultaneous presence of 0·1 μg ml−1 TTX, 20 mm TEA-Cl and 10 mm 4-AP (to block sustained and transient K+ currents as well as voltage-dependent Na+ channels). To prevent precipitation of the Ba2+, the medium was made devoid of all anions except Cl. All electrophysiological experiments were conducted at 31–33°C.

Measurements of glucagon release

Glucagon release was measured at +37°C following static incubations. Groups of 10 size-matched NMRI mouse islets were cultured overnight in RPMI-1640 medium and pre-incubated for 30 min in 200 μl extracellular solution consisting of (mm): 138 NaCl, 5·6 KCl, 2·6 CaCl2, 1·2 MgCl2, 5 Hepes (pH 7·4 with NaOH) and 0 or 10 D-glucose in 96-well Durapore membrane plates (Millipore, Molsheim, France). The pre-incubation medium was aspirated using a vacuum control pump (Millipore) and discarded. The islets were then resuspended in 200 μl of extracellular solution in the absence and presence of test compounds. At the end of the test incubation (1 h), the medium was aspirated and assayed immediately for glucagon using a commercial assay (GL-32K; Linco Research, St Charles, MO, USA).

Data analysis

All data are quoted as mean values ±s.e.m. Statistical significances were evaluated using Student's t test.


  1. Top of page
  2. Abstract
  6. Acknowledgements

Expression of voltage-gated Na+ currents in α-cells

Figure 1A shows Na+ currents evoked by voltage-clamp depolarisations to −40, −30, −20 and −10 mV in an α-cell within an intact mouse islet identified by the presence of a TEA-resistant K+ current and lack of glucose-induced electrical activity. Current responses were small at voltages below −30 mV. Depolarisations to more positive voltages evoked progressively larger inward currents. Figure 1B summarises the I–V relationship of the current recorded from 10 α-cells. The maximum peak current was attained between −10 and 0 mV reaching an average peak amplitude of 545 ± 103 pA. At more positive voltages, there was a secondary decline reflecting the decrease in driving force. The dotted curve represents the corresponding data recorded from δ-cells. It is clear that the Na+ currents in α-cells are 1·6-fold larger than those of δ-cells. Only part of this difference can be accounted for by differences in cell size as suggested by similar values for the cell capacitance (5·0 pF for α-cells and 4·4 pF for δ-cells; Göpel et al. 2000).


Figure 1. Characterisation of voltage-gated Na+ currents in pancreatic α-cells

A, family of voltage-clamp currents observed when the voltage was stepped from −70 mV to potentials between −40 and +50 mV in 10 mV increments. For clarity, only responses to depolarisations between −40 and −10 mV are shown. The depolarisations were 5 ms long and applied at a frequency of ≈1 Hz. The currents were recorded in the simultaneous presence of 20 mm TEA and 10 mm 4-AP to maximally reduce the outward K+ currents. B, I–V relationship of the Na+ current in α-cells. The dotted line shows the corresponding values in δ-cells (i.e. cells lacking an A-current). Data are mean values ±s.e.m. of 10 experiments. C, membrane currents recorded in α-cell in the presence of 20 mm TEA during 5 ms voltage-clamp depolarisations to −10 and 0 mV before and after addition of 0·1 μg ml−1 TTX. Note that the net current was outward at all times for both depolarisations in the presence of TTX. D, steady-state inactivation of the Na+ current in pancreatic α-cells. The cells were subjected to a conditioning prepulse (100 ms) to voltages between −120 and −10 mV prior to the 5 ms test pulse, which was at −10 mV. The cell was held at −70 mV for 1 ms between the conditioning pulse and the test pulse. The current elicited by the test pulse (I) following the conditioning prepulse to various membrane potentials (Vm) is plotted as a fraction (h) of the maximum current (Imax). Imax was measured following a prepulse to −110 mV. The curve was obtained by fitting the Boltzmann equation to the data points. Mean values ±s.e.m. of 4 experiments. The dotted curve represents the h-Vm curve for the Na+ current in the δ-cells.

Download figure to PowerPoint

As expected, the Na+ current in the α-cells can be blocked by tetrodotoxin (TTX; 0·1 μg ml−1, Fig. 1C). It is worthy of note that because of the presence of a rapidly activating K+ current (see Fig. 2 in the preceding paper and below), depolarisations applied in the presence of the Na+ channel blocker result in net outward currents at both −10 and 0 mV during the entire voltage pulse even when the experiment was conducted in the presence of TEA. When the Na+ channels are blocked by TTX, remaining inward conductances are too small to give rise to a net inward current at any time during depolarisation and consequently there is nothing to drive the upstroke of an action potential. Blockade of the Na+ channels with TTX can therefore be expected to strongly influence glucagon secretion (see below).


Figure 2. Properties of TEA-resistant K+ current (A-current) in α-cells

A, currents (top) were evoked by 200 ms depolarisations from the holding potential (−70 mV) to membrane potentials between −50 and −10 mV as indicated by the voltage protocol (bottom). The inactivation of the current could be described by a single exponential with a time constant τ. B, relationship between membrane potential during depolarisation and peak current amplitude (▴) and τ (▪). Data are mean values ±s.e.m. of 7 experiments.

Download figure to PowerPoint

Voltage-gated Na+ currents characteristically exhibit steady-state inactivation (Hille, 1992). To determine whether this also applies to the Na+ current in α-cells, we applied the traditional two-pulse protocol. The relationship between the voltage during the conditioning prepulse (V) and the relative peak current amplitude (h) in α-cells is summarised in Fig. 1D. The Boltzmann equation (eqn (1) in the preceding paper) was fitted to the data points to derive the inactivation parameters. In four α-cells, the values of Vh and kh averaged −47 ± 1 mV and −8 ± 1 mV, respectively. The continuous curve summarises the inactivation behaviour of the Na+ current in the α-cells. It can be observed that Na+ current inactivation occurs at voltages ∼20 mV more negative in the α-cells than in the δ-cells (dotted line).

Characterisation of TEA-resistant transient K+ current in pancreatic α-cells (‘A-current’)

In the preceding paper we demonstrated that δ- and α-cells contain transient outward K+ currents that differ with regard to the voltage dependence of activation, inactivation behaviour and TEA sensitivity (Göpel et al. 2000). Figure 2A shows voltage-gated K+ currents in an α-cell exposed to 20 mm TEA. The I–V relationship of the TEA-resistant K+ current component is summarised in Fig. 2B (▴). Outward current first became detectable during depolarisation to −40 mV. The peak current amplitude then increased progressively with the applied voltage. It is also clear that the current inactivated during the depolarisation. The time constant of inactivation (τ) amounted to 7 ± 1 ms (n= 7) at −40 mV and increased as with the applied voltage (Fig. 2B, ▪).

The inactivation properties of the TEA-resistant A-current were investigated using a standard two-pulse protocol. Figure 3A shows the current responses evoked by a 200 ms depolarisation to −10 mV following conditioning prepulses to voltages between −80 and −50 mV illustrating that inactivation occurs at rather negative membrane potentials. These experiments are summarised in Fig. 3B. In 14 cells, we obtained values for Vh and k of −68 ± 1 mV and −4 ± 1 mV, respectively. We acknowledge that > 30 % of the TEA-resistant current does not inactivate. One possibility is that the TEA-resistant current measured at −10 mV flows through more than one type of K+ channel, one of which does not undergo voltage-dependent inactivation.


Figure 3. Voltage-dependent inactivation of transient outward current

A (top), transient K+ currents in a functionally identified α-cell recorded in the presence of 20 mm TEA. A (bottom), prior to a test pulse to −10 mV, the α-cell was subjected to a 200 ms conditioning pulse to voltages between −100 and −10 mV. For clarity, only traces following pulses between −80 and −50 mV are shown. The test pulse was preceded by an interval of 20 ms during which the cell was held at −70 mV as indicated schematically by the voltage protocol. Each sequence was repeated at a frequency of 0·4 Hz. B, steady-state inactivation of the A-current. The current elicited by the test pulse (I) following conditioning prepulses to various membrane potentials (Vm) is plotted as a fraction (h) of the maximum current (Imax). Imax was measured following a prepulse to −100 mV. The curve was obtained by fitting the Boltzmann equation to the data points. Data are presented as mean values ±s.e.m. of 8 experiments. C, inhibition of TEA-resistant current by 10 mm 4-AP.

Download figure to PowerPoint

‘A-type’ K+ currents are characteristically resistant to TEA but sensitive to 4-aminopyridine (4-AP) (Hille, 1992; Conley, 1999). As shown in Fig. 3C, this also applies to the pancreatic variety of the current and when applied at a concentration of 10 mm, 4-AP reversibly abolished the transient outward current. In a series of four experiments, 10 mm 4-AP reduced the peak current by 66 ± 4 %. Thus, the IC50 (the concentration producing 50 % inhibition) is < 10 mm as expected for Kv4 channels (Conley, 1999). It is also evident that application of 4-AP not only reduces the amplitude of the A-current but also slows both its activation and inactivation resulting in a crossover of the currents recorded in the absence and presence of the blocker. This phenomenon has been reported in other cells and was attributed to 4-AP inhibiting the channel through a closed-state binding mechanism whereas the blocker does not interact strongly with the channel in its open or inactivated states (Campbell et al. 1993).

Voltage-gated Ca2+ current in the α-cells

The Ca2+ currents are not readily studied in the pancreatic α-cells using experimental protocols with K+-filled electrodes, a procedure necessary for the detection of the TEA-resistant K+ current. As shown in Fig. 1C, the rapid activation of the TEA-resistant A-current completely obscures any inward TTX-resistant currents. Moreover, the block by 4-AP leaves a sizeable part of the A-current (Fig. 3C). We therefore recorded α-cell Ca2+ current using Ba2+ as the charge carrier to block remaining voltage-gated K+ conductances. Figure 4A contains voltage-clamp Ba2+ currents flowing through α-cell Ca2+ channels elicited by voltage-clamp depolarisation from −70 mV to membrane potentials between −40 and −10 mV. Inward currents became detectable already during depolarisations to −40 mV, which elicited a slowly developing and inactivating current. Depolarisations to more positive voltages increased the rate of activation and inactivation of this current. An additional sustained component became apparent during depolarisations to membrane voltages beyond −20 mV. Both the inactivating and sustained components were blocked by addition of 5 mm Co2+ as expected for Ba2+ currents flowing through voltage-gated Ca2+ channels (Fig. 4B).


Figure 4. Two types of Ca2+ current in mouse α-cells

A, membrane currents recorded in the presence of 0·1 μg ml−1 TTX, 20 mm TEA and 4 mm 4-AP after equimolar substitution of Ba2+ for Ca2+ (2·6 mm). The currents evoked by depolarisations from −70 mV to the voltages indicated below the current traces are shown. B, inhibition of both Ca2+ current (top) components by Co2+ (5 mm) evoked by a voltage-clamp depolarisation from −70 to 0 mV (bottom). C, I–V relationships for the peak (▪) and sustained (□) Ba2+ currents flowing through α-cell Ca2+ channels measured as indicated schematically by the inset, which contain the currents evoked by a depolarisation to −10 mV. Data are given as mean values ±s.e.m. of 5 experiments. D, Ca2+ currents recorded during voltage ramps (rate, 0·35 V s−1) between −60 and +50 mV in α-cells (grey trace) and δ-cells (black trace). Note shoulder on α-cell trace that is attributable to T-type Ca2+ current and that is not present on δ-cell trace.

Download figure to PowerPoint

The I–V relationships of the α-cell Ca2+ currents recorded with Ba2+ as the charge carrier are shown in Fig. 4C. It can be observed that an inward current became detectable during depolarisations to −40 mV and beyond and the current amplitude (▪, see inset) amounted to −27 ± 3 pA (n= 5) at −30 mV. The maximum peak current was observed during depolarisations to 0 mV where it averaged −83 ± 9 pA (n= 5). For comparison the I–V relationship of the sustained current is also present (Fig. 4C; □). The sustained current became detectable during depolarisations to −20 mV and reached a maximum at voltages around 0 mV.

The presence of a Ca2+ current component activating at more negative voltages distinguishes the α-cell from the δ-cell (Göpel et al. 2000). This difference is highlighted in Fig. 4D in which the current responses to voltage ramps from −60 mV to +50 mV were applied. Whereas a clear shoulder is apparent in the α-cells at voltages between −50 and −20 mV with a secondary acceleration at voltages around −10 mV, the current amplitude in the δ-cells is negligible at membrane potentials below −30 mV. We attribute the two phases of the I–V relationship in the α-cell to the presence of both low- and high-threshold Ca2+ channels (i.e. T- and L-type), whereas the δ-cell only contains the high-threshold (L-type) variety. It is also notable that the high-threshold component reached its maximum at slightly more depolarised voltages in the α-cells than in the δ-cells and that the peak was narrower.

The Ba2+ currents could be described assuming m2h kinetics. The results are summarised in Fig. 5A. The continuous curves superimposed on the current traces show that this model faithfully predicts the activation and inactivation of the Ca2+ current both at −30 and at 0 mV. The time constant of activation fell from > 5 ms at −30 mV to 1 ms at voltages beyond 0 mV (Fig. 5B, ▪). The time constant of inactivation showed bell-shaped voltage dependence, being maximal during depolarisations between −20 and −10 mV and dropping to < 2 ms at +20 mV (Fig. 5B, •). At the latter voltages, the non-inactivating current predominated.


Figure 5. Ca2+ current in α-cells can be described by m2h kinetics

A, Ca2+ current evoked by membrane depolarisations to −30 and 0 mV. The superimposed curves represent m2h approximations to the currents (see Göpel et al. 2000). The values of τm and τh were 8·7 ms and 12·7 ms at −30 mV and 1·3 ms and 5·2 ms at 0 mV. In the latter case, 64 % of the current was taken to be non-inactivating. B, relationship between values of τm (▪) and τh (•) and membrane potential (V).

Download figure to PowerPoint

Steady-state inactivation of α-cell Ca2+ current

The voltage dependence of T- and L-type Ca2+ current inactivation was investigated by a voltage protocol similar to that used for the study of A-current inactivation (Fig. 3A). Figure 6A illustrates the steady-state inactivation of the T-type Ca2+ current. In these experiments, a test pulse to −30 mV was preceded by conditioning prepulses to membrane potentials between −80 and −10 mV (−60 to −30 mV shown). The results are summarised in Fig. 6B. The current amplitude was unaffected by varying the conditioning voltage in the range −110 and −60 mV. At more depolarised voltages, the amplitude of the transient current gradually declined and following conditioning pulses to voltages beyond −30 mV it was fully inactivated. In five different cells, the values of Vh and k averaged −45 ± 1 mV and −4 ± 1 mV, respectively.


Figure 6. Inactivation properties of Ca2+ currents in α-cells

A (bottom), a 200 ms test pulse to −30 mV was preceded by 200 ms conditioning pulses to membrane potentials between −100 and −10 mV. For clarity, only responses following pulses between −60 and −30 mV are shown. The cell was held at −70 mV for 10 ms between the conditioning prepulse and the test pulse. A (top), currents measured following conditioning pulses to (from bottom to top) −60, −50, −40 and −30 mV. B, steady-state inactivation of T-type Ca2+ current in an α-cell. The current elicited by the test pulse (I) following conditioning prepulses to various membrane potentials (Vm) is plotted as a fraction (h) of the maximum current (Imax). Imax was measured following prepulses to membrane potentials <=−90 mV. The curve was obtained by fitting the Boltzmann equation to the data points. Mean values ±s.e.m. of 5 experiments. C, as in A except that the test pulse was −20 mV to activate both the inactivating and the sustained components. D, as in B but the relative amplitude (h=I/Imax) of the sustained current component, measured at the end of the test depolarisation, is shown. Mean values ±s.e.m. of 4 experiments.

Download figure to PowerPoint

Figure 6C shows the current responses in an experiment in which the test pulse went to −20 mV to activate both the sustained L-type and the transient T-type components. When the conditioning voltage was ≤−50 mV, both the sustained and transient components were activated by the subsequent test depolarisation. However, following a conditioning pulse to −20 mV, the current elicited by the test depolarisation consisted solely of the sustained component. The h-Vm relationship of the sustained component (measured at the end of the 200 ms test depolarisation) is shown in Fig. 6D. The sustained Ca2+ current component is unaffected by the conditioning voltage, at least when Ba2+ was used as the charge carrier. The latter current is likely to reflect the pure high-threshold (L-type) component. The activation of this current was described by m2 kinetics yielding a time constant of 1·4 ± 0·2 ms (n= 4) at −20 mV.

Regenerative α-cell electrical activity

As shown in the accompanying paper (Göpel et al. 2000), electrical activity in α-cells can occasionally be evoked by hypoglycaemia. More often, α-cells generated just a few action potentials following the release of the voltage clamp and then settled at a fairly depolarised membrane potential (Fig. 7A). Under these conditions, action potentials could be elicited by current injection. Figure 7B shows changes in membrane potential evoked by injecting currents (−7 to −1 pA in 2 pA increments) in an α-cell held at ∼-80 mV by injection of −9 pA hyperpolarising current. It can be seen that injection of small currents exerted marked effects on the α-cell membrane potential. Once the stimulus became sufficient, large and overshooting action potentials could be elicited.


Figure 7. Action potentials in α-cells

A, electrical activity recorded in an α-cell exposed to 1 mm glucose upon switching the amplifier from the voltage-clamp to the current-clamp mode (*). Prior to the release of the voltage clamp, the membrane potential was held at −70 mV by injection of −4 pA current. B, changes of α-cell membrane potential when applying current steps (2 to 8 pA) in an α-cell held at a membrane potential more negative than −80 mV by injection of −9 pA of current.

Download figure to PowerPoint

Regulation of α-cell electrical activity by the KATP channel blocker tolbutamide

KATP channels have previously been documented in rat α-cells (Bokvist et al. 1999). Here we have investigated whether these channels are also present in mouse α-cells and if they are capable of modulating the electrical activity in these cells. Figure 8A shows a recording of the membrane potential in an α-cell. The experiment commenced in the presence of 10 mm glucose. The membrane potential of α-cells under these conditions was, contrary to our expectation, fairly depolarised and averaged −42 ± 6 mV (n= 7). However, a spontaneous and gradual hyperpolarisation and increase in membrane conductance was frequently observed within a few minutes of establishing the perforated patch whole-cell configuration irrespective of whether the recording was made in the absence or presence of glucose. For example, the α-cell in Fig. 8A repolarised by ∼30 mV during the first 90 s after establishing the perforated patch whole-cell configuration. In four different cells, the membrane potential eventually stabilised at −76 ± 1 mV. The reason for this hyperpolarisation is unknown but it is possible that merely establishing perforated patch recording conditions interferes sufficiently with α-cell metabolism to produce activation of K+ channels.


Figure 8. KATP channel activity influences the membrane potential of the α-cell

A, membrane potential recording from an α-cell. The recording commenced in the presence of 10 mm glucose and the membrane potential was −35 mV. The cells then repolarised spontaneously and 30 min later (*) the membrane potential had settled at −80 mV. Tolbutamide (0·1 mm) was then applied during the period indicated by the bar. At the times indicated by a and b, the recording was interrupted, the amplifier was switched into the voltage-clamp mode and ±10 mV pulses applied from −70 mV to monitor membrane conductance. Note that the effect of tolbutamide was reversed upon its withdrawal from the medium and that repolarisation led to brief reappearance of electrical activity. B, membrane currents evoked by 10 mV hyperpolarising pulses in the absence (a) and presence (b) of 0·1 mm tolbutamide. C, changes of membrane potential occurring during application of tolbutamide in A shown on an expanded time base. Note that the first action potential is fired when the action potential is ≈−60 mV and that action potential firing ceased when the membrane potential exceeded −40 mV (dotted horizontal lines).

Download figure to PowerPoint

Addition of tolbutamide (0·1 mm) to α-cells that had spontaneously hyperpolarised resulted in strong membrane depolarisation and induced a brief period of electrical activity before the membrane potential settled at a new value of −33 ± 4 mV (P < 0·01). However, once the plateau potential had been attained, only weak and irregular membrane potential oscillations were observed which were reminiscent of those observed in the presence of 10 mm glucose (compare Fig. 1D in the accompanying paper). As illustrated in Fig. 8B, the induction of electrical activity was associated with a decrease in membrane conductance from 1·08 ± 0·11 nS under control conditions to 0·38 ± 0·06 nS (n= 4; P < 0·01) in the presence of tolbutamide. By contrast, neither the membrane potential nor the membrane conductance of the α-cell was affected by glucose under these experimental conditions. When tolbutamide was subsequently removed, the α-cell repolarised towards the pre-stimulatory membrane potential. Interestingly, repolarisation was associated with the transient reappearance of action potential firing suggesting that some (low) KATP channel activity is required for regenerative electrical activity in the α-cell. Figure 8C shows that the first action potential is elicited when the membrane potential exceeds −60 mV and the last spike is evoked from a membrane potential of ∼-35 mV. Our value for the whole-cell KATP conductance in the α-cell is only 10 % of that observed in the β-cell under the same experimental conditions (Göpel et al. 2000). The low KATP channel density is in accordance with the reported weak binding of fluorescently labelled glibenclamide to dispersed α-cells maintained in tissue culture (Quesada et al. 1999).

Glucagon secretion

We finally correlated the electrophysiological findings to modulation of glucagon release (Table 1). As expected, increasing glucose from 0 to 10 mm resulted in a 47 % reduction of glucagon release. Addition of the Ca2+ channel blocker nifedipine reduced glucagon release to the same extent as glucose. Importantly, the Na+ channel inhibitor TTX was a stronger inhibitor of glucagon release than glucose itself. The KATP channel activator diazoxide reduced glucagon release to the same extent as glucose. Surprisingly, but in keeping with the data of Fig. 9, tolbutamide exerted a stronger inhibitory action than 10 mm glucose and was in fact as effective as TTX in suppressing glucagon release.

Table 1.  Influence of ion channel modulators on glucagon release
ConditionGlucagon release (pg (10 islets)−1 h−1)
  1. Glucagon release was measured following 1 h incubations under the indicated experimental conditions. Unless otherwise indicated, glucagon release was measured in the absence of glucose. Data are mean values ±s.e.m. of indicated (n) number of measurements. *P < 0.001vs. control.

Control (0 mm glucose)3182 ± 171 (n= 15)
25 μm nifedipine1793 ± 168* (n= 10)
1 μg ml−1 TTX1176 ± 70* (n= 10)
100 μm tolbutamide1275 ± 132* (n= 15)
100 μm diazoxide1690 ± 219* (n= 10)
10 mm glucose1700 ± 120* (n= 20)

Figure 9. Bell-shaped relationship between concentration of diazoxide and glucagon secretion

Intact islets were exposed to 10 mm glucose. Diazoxide was added at concentrations between 1 and 100 μm. The shaded area indicates glucagon release in the absence of glucose. Data are presented as mean values ±s.e.m. of 5–8 experiments. *P < 0·05; **P < 0·01.

Download figure to PowerPoint

In a second series of experiments we tested the hypothesis that reactivation of KATP channels could stimulate glucagon release as suggested by the period of electrical activity observed upon withdrawal of tolbutamide (Fig. 8A). Figure 9 shows the effects of diazoxide at concentrations between 0 and 100 μm on glucagon secretion in the presence of 10 mm glucose. It is clear that the two lowest concentrations of diazoxide (1 and 2 μm) stimulated glucagon secretion to the same extent as that obtained in response to a lowering of glucose from 10 to 0 mm (shaded area). Higher concentrations of diazoxide had either a weaker stimulatory effect (5 μm) or no effect at all (25 μm) whereas 100 μm in fact produced stronger inhibition of glucagon release than glucose itself (compare with value for zero diazoxide). For comparison, the half-maximal stimulatory effect of diazoxide on KATP channel activity in β-cells has been reported to occur at ∼80 μm (Schwanstecher et al. 1992).


  1. Top of page
  2. Abstract
  6. Acknowledgements

In this and the accompanying paper we make use of a method that allows patch-clamp experiments to be performed on the endocrine cells of intact mouse pancreatic islets. We have previously characterised the membrane conductances of β-cells (Göpel et al. 1999a, b) and δ-cells in situ (Göpel et al. 2000). In this paper we focus on the electrophysiology of the α-cell. Based on our findings we propose a novel hypothesis for the regulation of glucagon secretion by glucose.

Na+ channels are important for α-cell electrical activity and glucagon release

Like δ-cells, α-cells are equipped with TTX-sensitive Na+ channels. However, close inspection reveals at least two important differences between the Na+ current in α- and δ-cells. First, the Na+ current of the α-cells is about 60 % larger than that of the δ-cell. Second, inactivation of the Na+ current occurs at more negative voltages in α-cells than in δ-cells (Vh=−47 mV in α-cells vs.−28 mV in δ-cells; Fig. 1D). It is of interest that Plant (1988) reports the presence of Na+ currents with inactivation occurring at physiological membrane potentials (Vh=−50 mV) in isolated islet cells that were not electrically active in the presence of glucose. Given the present observations it seems likely that these cells were in fact α-cells. As will be discussed below, inactivation of the Na+ current at fairly negative voltages has important functional implications and may, for example, explain the divergent effects of glucose on glucagon and somatostatin secretion.

The α-cells contain a TEA-resistant A-current

Both δ- and α-cells are equipped with a transient inactivating K+ current. Again there are important differences between the two cell types. In the α-cells, the current both activates (threshold −50 mV vs.−30 mV) and inactivates (Vh at −69 mV vs.−34 mV) at more negative voltages than in the α-cells. The inactivating K+ current in the α-cell, but not that in the δ-cell, is resistant to TEA and sensitive to 4-AP, thus conforming to the pharmacological profile of archetypal A-type K+ channels. A TEA-resistant current with similar inactivation properties has previously been observed in dispersed pancreatic islet cells (Smith et al. 1989). The inactivation properties of the latter current (Vh at −72 mV) closely resemble those of the transient K+ current we now report in the α-cells (−69 mV). Given the data presented here it seems likely that the A-currents described in the earlier study were recorded from α-cells. This interpretation is consistent with the reported low value of the cell capacitance of these cells (3·3 pF; Smith et al. 1989).

β-cells and α-cells are not electrically coupled

We have reported elsewhere that whereas β-cells were electrically coupled to each other, there was no electrophysiological evidence for coupling between β-cells and non-β-cells (Göpel et al. 1999a). The same conclusion has recently been reached by comparing the [Ca2+]i oscillations in α-, β- and δ-cells in intact islets (Nadal et al. 1999) but is contrary to the early proposal by Meda et al. (1986). However, given the electrophysiological characteristics of the α-cells, lack of electrical coupling between β- and α-cells is a sensible functional adaptation. We show here that injection of currents as small as a few picoamps is sufficient to evoke electrical activity (Fig. 7). If α-cells were electrically coupled to β-cells and assuming the same gap-junctional conductance as between β-cells (1 nS; Göpel et al. 1999a), then glucose-induced depolarisation of the β-cell (10–30 mV) would inject depolarising currents as large as 10–30 pA into the α-cell. This would clearly be sufficient to evoke action potentials in the α-cell and thus stimulation of glucagon release. These considerations suggest that the electrical coupling between α- and β-cells must be weak (if not lacking entirely) to allow glucagon and insulin release to be regulated reciprocally by glucose.

Mouse pancreatic α-cells contain two types of Ca2+ channel

As has previously been demonstrated in guinea-pig α-cells (Rorsman, 1988), mouse α-cells are equipped with two types of Ca2+ channel with functional characteristics similar to T- and L-type Ca2+ channels in other cell types (Bean, 1989). The low-threshold component (T-type) becomes detectable already during depolarisations to voltages more negative than −40 mV and is subject to steady-state inactivation (Vh=−45 mV). The sustained (L-type) Ca2+ current first becomes detectable during depolarisations to −20 mV. In terms of Ca2+ channel density, the maximum current observed in the α-cells with Ba2+ as the charge carrier is ∼80 pA at 0 mV, which is somewhat less than the corresponding value in δ-cells (100 pA; Göpel et al. 2000). Comparing the Ba2+ currents evoked by the voltage ramps (Fig. 4D) further suggests that the L-type Ca2+ current in α-cells requires slightly more depolarised voltages for activation than that of the δ-cells. These features support the argument that the L-type Ca2+ current may not be sufficient to sustain electrical activity and secretion itself. Rather, glucagon release requires large depolarisations to maximally activate the L-type Ca2+ current and this role is fulfilled by the rapid activation of the Na+ channels during the upstroke of the α-cell action potential. This scenario is fully compatible with the strong inhibitory action of TTX on glucagon release.

Electrical activity in α-cells

Comparing electrical activity in mouse α-cells with membrane potential recordings from δ- and β-cells reveals a number of interesting differences. The α-cells are unique among the islet cells in generating overshooting action potentials that originate from a negative membrane potential (−60 mV; see Fig. 8C). This range of membrane potentials is close to that where the T-type Ca2+ channels start activating and suggests that the T-type Ca2+ current, by analogy to the situation documented in other cells (Perez-Reyes et al. 1998), serves a pacemaker function in the α-cell. A similar conclusion has previously been reached for guinea-pig α-cells (Rorsman, 1988). The activation of T-type Ca2+ channels brings the membrane potential of the α-cell into the range where the Na+ channels and L-type Ca2+ channels begin to open (between −30 and −20 mV). Repolarisation of the action potential probably results from the fast activation of the A-current. From these considerations it follows that conditions resulting in the inactivation of the T-type Ca2+ current will strongly influence the electrical and secretory activity of the α-cell. The steady-state inactivation properties of the T-type Ca2+ current are such that it is almost fully inactivated at voltages beyond −40 mV, a range of membrane potentials that is too negative for activation of the Na+ channels and the L-type Ca2+ channels.

Effects of sulphonylureas in the α-cell

Our studies have revealed that β-cells (Göpel et al. 1999a, b), δ-cells (Göpel et al. 2000) and α-cells (this study) in situ are all equipped with KATP channels. Whereas the role of these channels in the δ-cells is straightforward and the same as in the β-cell (to couple the cell's metabolic state to changes of the membrane potential), the functional significance of such channels in the α-cells is less obvious. At first sight, regulation of these channels by glucose appears counter-intuitive as this would result in membrane depolarisation and thus, one could surmise, stimulation of electrical and secretory activity. However, the experiments with tolbutamide suggest a different scenario. We demonstrate here that addition of tolbutamide leads to a strong depolarisation and a transient stimulation of electrical activity. Once the depolarised plateau (−30 mV) had been attained, however, only small and irregular oscillations in the membrane potential were observed. The absence of regenerative electrical activity on the plateau we attribute to the inactivation of the Na+ current, T-type Ca2+ current and A-current, all of which are almost completely inactivated at voltages beyond −50 mV. The voltage at the plateau is below that required for opening of the L-type Ca2+ channels (Fig. 4C and D). Accordingly, closure of the KATP channels and membrane depolarisation in the α-cell (unlike the situation in the β- and δ-cell) may be coupled to inhibition rather than stimulation of glucagon release.

The paradox that both tolbutamide and diazoxide inhibit glucagon release can be reconciled with this model. Tolbutamide inhibits glucagon release by causing inactivation of the voltage-gated membrane conductances as discussed above. Diazoxide leads to strong activation of the KATP channels and thus hyperpolarises the α-cell. In both cases, action potential firing ceases and glucagon secretion is suppressed but the mechanisms involved are completely different.

A possible model for regulation of glucagon release by glucose

The scenario for regulation of glucagon release by tolbutamide we outline here may also help to explain the mechanism by which glucose inhibits glucagon secretion. Glucose-induced inhibition of glucagon release is dependent on glucose metabolism (Östenson, 1980; Heimberg et al. 1995). Measurements of the adenine nucleotide content in purified rat α-cells have revealed that they have a high (∼7) ATP/ADP ratio already at low (1 mm) glucose and that it changes little (if at all) during glucose stimulation (Detimary et al. 1998). This contrasts to the situation in β-cells where the ATP/ADP ratio increases from 2·5 at 1 mm glucose to > 12 at glucose concentrations ≥ 10 mm (Detimary et al. 1998). It is of interest that the ATP/ADP ratio in rat α-cells at 1 mm glucose is about the same as in β-cells exposed to 6 mm glucose. The latter concentration is associated with strong inhibition of the KATP channels and sets the membrane potential just below the threshold for initiation of electrical activity in β-cells. If these data can be extrapolated to mouse α-cells, then the KATP channel activity during hypoglycaemia can be expected to be low (but greater than zero) and the membrane potential close to that from which the action potentials originate.

Based on the observations made with tolbutamide, we propose that glucose suppresses glucagon secretion by inducing closure of remaining KATP channels leading to membrane depolarisation and inactivation of the ion channels participating in action potential generation. The metabolic product of glucose that links glucose metabolism to KATP channel inhibition in the α-cell remains to be determined. In view of the experiments on purified rat α-cells described above, it seems unlikely to be ATP as neither the total concentration of this nucleotide nor the ATP/ADP ratio changes much upon glucose stimulation. However, it may be worth considering GDP as a coupling factor as this nucleotide is as effective as ADP in activating KATP channels closed by ATP (Bokvist et al. 1999).

In this scenario the KATP channels serve exactly the same function in the β- and α-cell and elevation of glucose leads to membrane depolarisation in both cell types. The opposite effects of glucose and tolbutamide on glucagon- and insulin secretion is then the consequence of the cells being equipped with different sets of voltage-dependent ion channels. Thus, whereas electrical activity in the β-cell depends on the activation of L-type Ca2+ channels and delayed rectifying K+ channels, neither of which undergoes voltage-dependent inactivation (Ashcroft & Rorsman, 1989), electrical activity in the α-cells is attributable to the opening of voltage-gated Na+ channels, T-type Ca2+ channels and A-type K+ channels, all of which inactivate at voltages beyond −50 mV.

Our model predicts that action potential firing in the α-cell is only possible in a narrow window of KATP channel activity where the membrane potential is positive enough to allow opening of pacemaker ion channels (T-type Ca2+ channels?) whilst at the same time being sufficiently negative to prevent voltage-dependent inactivation of the membrane conductances involved in action potential generation. The ability of low concentrations of diazoxide to stimulate glucagon release in α-cells exposed to 10 mm glucose supports this hypothesis. This model for regulation of electrical activity and glucagon secretion in α-cells also explains the reported paradoxical observation that application of diazoxide produced a slight reduction of basal [Ca2+]i and reappearance of [Ca2+]i oscillations in some α-cells in intact islets exposed to high glucose (Quesada et al. 1999). In view of our findings, it seems likely that the latter effect resulted from the opening of a few KATP channels with a resultant increase in membrane potential and recovery from voltage-dependent inactivation. Given the important role KATP channels may play in the regulation of glucagon secretion it is now essential to study the metabolic regulation of the α-cell KATP channels in greater detail.

  • Ashcroft, F. M. & Rorsman, P. (1989). Electrophysiology of the pancreatic β-cell. Progress in Biophysics and Molecular Biology 54, 87143.
  • Bean, B. P. (1989). Classes of calcium channels in vertebrate cells. Annual Review of Physiology 51, 367384.
  • Berts, A., Ball, A., Gylfe, E. & Hellman, B. (1996). Suppression of Ca2+ oscillations in glucagon-producing α2-cells by insulin/glucose and amino acids. Biochimica et Biophysica Acta 1310, 212216.
  • Bokvist, K., Olsen, L. H., Høy, M., Gotfredsen, C. F., Holmes, W. F., Buschard, K., Rorsman, P. & Gromada, J. (1999). Characterisation of sulphonylurea and ATP-regulated K+-channels in rat pancreatic A-cells. Pflügers Archiv 438, 428436.
  • Campbell, D. L., Qu, Y., Rasmusson, R. L. & Strauss, H. C. (1993). The calcium-independent transient outward potassium current in isolated ferret right ventricular myocytes. II. Closed state reverse use-dependent block by 4-aminopyridine. Journal of General Physiology 101, 603626.
  • Conley, E. C. (1999). VLG Kv4-Shal. In The Ion Channel Facts Book: Voltage-gated channels, vol. 4, ed. Conley, E. C. & Brammar, W. J., pp. 617646. Academic Press, San Diego .
  • Detimary, P., Dejonghe, S., Ling, Z., Pipeleers, D., Schuit, F. & Henquin, J. C. (1998). The changes in adenine nucleotides in glucose-stimulated rodent islets occur in β cells but not in α cells and are also observed in human islets. Journal of Biological Chemistry 273, 3390533908.
  • Göpel, S., Kanno, T., Barg, S., Galvanovskis, J. & Rorsman, P. (1999a). Voltage-gated and resting membrane currents recorded from B-cells in intact pancreatic islets. The Journal of Physiology 521, 717728.
  • Göpel, S. O., Kanno, T., Barg, S., Eliasson, L. J, Galvanovskis, J., Renström, E. & Rorsman, P. (1999b). Activation of Ca2+-dependent K+ channels contributes to rhythmic firing of action potentials in mouse pancreatic β cells. Journal of General Physiology 114, 759769.
  • Göpel, S. O., Kanno, T., Barg, S. & Rorsman, P. (2000). Patch-clamp characterisation of somatostatin-secreting δ-cells in intact mouse pancreatic islets. The Journal of Physiology 528, 497507.
  • Gromada, J., Bokvist, K., Ding, W. G., Barg, S., Buschard, K., Renström, E. & Rorsman, P. (1997). Adrenaline stimulates glucagon secretion in pancreatic A-cells by increasing the Ca2+ current and the number of granules close to the L-type Ca2+ channels. Journal of General Physiology 110, 217228.
  • Heimberg, H., De vos, A., Pipellers, D., Thorens, B. & Schuit, F. (1995). Differences in glucose transporter gene expression between rat pancreatic alpha- and beta-cells are correlated to differences in glucose transport but not in glucose utilization. Journal of Biological Chemistry 270, 89718975.
  • Hille, B. (1992). Ion Channels of Excitable Membranes. Sinauer, Sunderland , MA , USA .
  • Meda, P., Santos, R. M. & Atwater, I. (1986). Direct identification of electrophysiologically monitored cells within intact mouse islets of Langerhans. Diabetes 35, 232236.
  • Nadal, A., Quesada, I. & Soria, B. (1999). Homologous and heterologous asynchronicity between identified α, β and δ cells within intact islets of Langerhans in the mouse. The Journal of Physiology 517, 8593.
  • ÖStenson, C.-G. (1980). Alloxan reversibly impairs glucagon release and glucose oxidation by pancreatic A2-cells. Biochemical Journal 188, 201206.
  • Perez-Reyes, E., Cribbs, L. L., Daud, A., Lacerda, A. E., Barclay, J., Williamson, M. P., Fox, M., Rees, M. & Lee, J. H. (1998). Molecular characterization of a neuronal low-voltage-activated T-type calcium channel. Nature 391, 896900.
  • Plant, T. D. (1988). Na+ currents in cultured mouse pancreatic B-cells. Pflügers Archiv 411, 429435.
  • Quesada, I., Nadal, A. & Soria, B. (1999). Different effects of tolbutamide and diazoxide in α-, β- and δ-cells within intact islets of Langerhans. Diabetes 48, 23902397.
  • Rorsman, P. (1988). Two types of Ca2+-current with different sensitivity of Ca2+-channel antagonists in guinea pig pancreatic α2 cells. Journal of General Physiology 91, 210222.
  • Rorsman, P. & Hellman, B. (1988). Voltage-activated currents in guinea pig pancreatic α2 cells. Evidence for Ca2+-dependent action potentials. Journal of General Physiology 91, 223242.
  • Schwanstecher, C., Dickel, C., Ebers, I., Lins, S., Zünkler, B. J. & Panten, U. (1992). Diazoxide-sensitivity of the adenosine 5′-triphosphate-dependent K+ channel in mouse pancreatic β-cells. British Journal of Pharmacology 107, 8794.
  • Smith, P. A., Bokvist, K. & Rorsman, P. (1989). Demonstration of A-currents in pancreatic islet cells. Pflügers Archiv 413, 441443.
  • Yoshimoto, Y., Fukuyama, Y., Horio, Y., Inanobe, A., Gotoh, M. & Kurachi, Y. (1999). Somatostatin induces hyperpolarization in pancreatic islet α cells by activating a G protein-gated K+ channel. FEBS Letters 444, 265269.


  1. Top of page
  2. Abstract
  6. Acknowledgements

This study was supported by the Swedish Diabetes Association, The Juvenile Diabetes Foundation International, the Knut and Alice Wallenbergs Stiftelse, the Swedish Medical Research Council (grants 8647, 12234 and 13147), the Crafoord Foundation, the Swedish Foundation for Strategic Research, the Aage and Louise Hansen Foundation, the Novo Nordisk Foundation and the Medical Faculty, Lund University. Dr Kanno's stay at Lund University was supported by Hirosaki University, Japan. Dr Weng is on temporary leave from the First Hospital of Xinxiang Medical College and supported by a CSC scholarship.

S. O. Göpel and T. Kanno contributed equally to this study and their names appear in alphabetical order.