Sensitivity of Kir6.2-SUR1 currents, in the absence and presence of sodium azide, to the K_ATP channel inhibitors, ciclazindol and englitazone

Abstract

1. Two electrode voltage clamp and single channel recordings were used to investigate the actions of various ATP-sensitive K⁺ (K_ATP) channel inhibitors on cloned K_ATP channels, expressed in Xenopus oocytes and HEK 293 cells.

2. Oocytes expressing Kir6.2 and SUR1 gave rise to inwardly rectifying K⁺ currents following bath application of 3 mM sodium azide. Inside-out recordings from non-azide treated oocytes demonstrated the presence of K_ATP channels which were activated by direct application of 3 mM azide and 0.1 mM Mg-ATP.

3. Tolbutamide inhibited azide-induced macroscopic Kir6.2-SUR1 currents, recorded from Xenopus oocytes, with an IC₅₀ value similar to native K_ATP channels. Ciclazindol and englitazone also inhibited these currents in a concentration-dependent manner, but with relative potencies substantially less than for native K_ATP channels.

4. Single channel currents recorded from inside-out patches excised from oocytes expressing Kir6.2-SUR1 currents were inhibited by tolbutamide, Mg-ATP, englitazone and ciclazindol, in the absence of azide, with potencies similar to native K_ATP channels. In the presence of azide, Kir6.2-SUR1 currents were inhibited by englitazone and tolbutamide but not ciclazindol.

5. Single channel currents derived from Kir6.2Δ26, expressed in HEK 293 cells, were inhibited by ciclazindol and englitazone irrespective of the absence or presence of SUR1.

6. In conclusion, heterologously expressed Kir6.2 and SUR1 recapitulate the pharmacological profile of native pancreatic β-cell K_ATP channels. However, currents induced by azide exhibit a substantially reduced sensitivity to ciclazindol. It is likely that ciclazindol and englitazone inhibit K_ATP currents by interaction with the Kir6.2 subunit.

Introduction

The electrical state of a cell can be coupled to its metabolic condition through modulation of ATP-sensitive potassium (K_ATP) channel activity. Consequently, these channels, which are present in many tissues, are considered to play an important role in both normal and diseased physiology including control of muscle function, neurological disorders and insulin secretion (Ashcroft & Ashcroft, 1990). Pharmacological manipulation of K_ATP channels is of obvious therapeutic value, clearly demonstrated in the treatment of hypo- and hyperglycaemia (Ashford, 1990). Sulphonylureas such as tolbutamide and glibenclamide are routinely used to treat non-insulin dependent diabetes mellitus (NIDDM) by stimulating insulin secretion, whereas the inhibition of insulin release by K_ATP openers such as diazoxide is sometimes used in the treatment of persistent hyperinsulinaemic hypoglycaemia of infancy (Dunne et al., 1997).

The molecular architecture of the pancreatic β cell K_ATP channel complex has been proposed to consist of at least two distinct proteins, a pore forming inwardly rectifying potassium channel sub-unit (Kir6.2) and an associated modulatory sulphonylurea receptor (SUR1). Current models propose that the K_ATP channel is an octomeric complex with four Kir6.2 sub-units, which constitute the channel pore, and four SUR1 sub-units which endow the channel with sensitivity to the sulphonylureas (Tucker & Ashcroft, 1998; Babenko et al., 1998). Heterologous expression of Kir6.2 and SUR1 in Xenopus oocytes, have demonstrated that expression of Kir6.2 or SUR1 alone does not give rise to functional K_ATP channels but when Kir6.2 and SUR1 are co-expressed, large inwardly rectifying currents are observed in response to application of metabolic inhibitors such as sodium azide (Gribble et al., 1997a).

The pharmacology of azide-induced currents appears to mimic that of native pancreatic β cell K_ATP channels as Kir6.2-SUR1 currents are inhibited by the sulphonylureas tolbutamide and glibenclamide and activated by diazoxide (Gribble et al., 1997a). Furthermore truncation of Kir6.2 by up to 36 carboxy-terminal amino acids induces channel expression in the absence of SUR1 and this has shown that the site of interaction for ATP is present on Kir6.2 alone, and that sensitivity to sulphonylureas, diazoxide and the potentiation of channel activity by Mg-ATP requires co-expression of the truncated Kir6.2 with SUR1 (Tucker et al., 1997). However, not all drugs require the presence of SUR1 to enable inhibition of the pore-forming subunit, Kir6.2, as the imidazolines (Proks & Ashcroft, 1997) and the antiarrhythmic agent, cibenzoline (Mukai et al., 1998) block Kir6.2 currents in the absence of SUR1.

Two other, structurally unrelated, agents have also recently been shown to inhibit β-cell-type K_ATP channel activity. Thiazolidinediones, used to enhance peripheral insulin sensitivity, such as troglitazone (Lee et al., 1996a) and englitazone (Rowe et al., 1997) and the putative anti-depressant and anorectic agent, ciclazindol (Lee et al., 1996b), block K_ATP channels in the insulin secreting cell line, CRI-G1. The inhibition of K_ATP channel currents by both englitazone and ciclazindol is unaffected by removal of intracellular Mg²⁺ ions and after trypsinization of the cytoplasmic surface of excised patches, treatments known to abolish sulphonylurea sensitivity (Lee et al., 1994a,1994b), indicating that these drugs may act at a site distinct from the sulphonylureas.

It is generally considered that sodium azide activates K_ATP currents by depleting intracellular ATP through uncoupling mitochondrial oxidative phosphorylation by inhibition of cytochrome oxidase (Tsubaki & Yoshikawa, 1993). However, recently it has been shown that sodium azide activates K_ATP channels in the absence of any metabolic-induced decline of ATP levels, indicating a more direct effect on this channel type (Harvey et al., 1999). Consequently, using whole-cell and single channel recordings of Kir6.2-SUR1 currents from Xenopus oocytes and HEK 293 cells we have examined whether azide directly activates Kir6.2-SUR1 channels and determined the sensitivity of the cloned K_ATP channel to the sulphonylureas, tolbutamide and glibenclamide in comparison to englitazone and ciclazindol, in the presence and absence of azide. In addition we have addressed the question as to which subunit is responsible for the actions of these agents.

Methods

Preparation of in vitro cRNA and cDNAs for transfection

cDNAs encoding mouse Kir6.2 and a C-terminal truncated form of Kir6.2, missing the last 26 amino acids, Kir6.2Δ26 (provided by Professor F. Ashcroft, University of Oxford), and rat SUR1 (provided by Dr G. Bell, University of Chicago) were used in this study. In vitro cRNA was prepared from Kir 6.2 and SUR1 cDNAs. One μg of linear template cDNA was used to prepare capped cRNA by in vitro transcription using the mMessage mMachine (Ambion Inc.) according to the manufacturer's instructions. The cRNAs were aliquoted and stored in DEPC-treated distilled water at −70°C until required. Endotoxin free Kir 6.2, SUR1 and Kir6.2Δ26 cDNAs for transfection was prepared using a Qiagen endotoxin free plasmid kit (Qiagen).

Oocyte collection and preparation

Ovarian lobes were removed from mature female Xenopus laevis frogs (Blades Biological) after killing the animal by destruction of the brain. Oocytes were separated by treatment with 1 U ml⁻¹ collagenase (Type 1A, Sigma) for 60–90 min at room temperature. Defollicated stage V–VI oocytes were selected for injection using a stereomicroscope on the basis of size, morphology and clear differentiation of animal and vegetal poles. Oocytes were injected either with 50 nl of a 1 : 1 mix of Kir6.2 and SUR1 cRNAs (∼1 ng of each cRNA) or 50 nl of water as a control. Isolated oocytes were maintained at 18°C in modified Barth's solution containing (mM): NaCl 110, KCl 1, NaHCO₃ 2.4, Tris-HCl 7.5, Ca(NO₃)₂ 0.41, CaCl₂ 0.41, MgSO₄ (pH 7.4) 0.82 supplemented with 50 mg ml⁻¹ gentamycin (Sigma) and 550 mg ml⁻¹ pyruvic acid (Sigma). Currents were studied 2–5 days post-injection.

Cell culture and transfection

HEK 293 cells were cultured in Dulbecco's modified Eagle's medium (Life Technologies) containing 1000 mg l⁻¹ glucose and 110 mg l⁻¹ sodium pyruvate with pyridoxine 4 mg l⁻¹ supplemented with 10% foetal bovine serum (Helena Biosciences) and 1% penicillin/streptomycin (Life Technologies). Cells were incubated at 37°C in a humidified atmosphere containing 5% CO₂ and plated onto 35 mm culture dishes (Falcon 3001) such that after 24 h the dishes were 40% confluent prior to transfection. Each dish was co-transfected with either 1.4 μg of a 1 : 1 ratio of Kir6.2Δ26/CD4 cDNAs or 3.15 μg of a 1 : 1 ratio of Kir6.2Δ26/SUR1/CD4 cDNAs using Fugene 6 transfection reagent (Roche Diagnostics Ltd.). A ratio of 2 : 3 μg cDNA to μl Fugene 6 was used as per the manufacturer's protocol. The culture media was replaced after 4 h and the cells were used for recording 24–48 h post-transfection. Transfected cells were selected by visible binding of anti-CD4 coated beads (Dynal, Oslo) following incubation with the beads for 20 min.

Drugs and solutions

Sodium azide, adenosine-5′-triphosphate (magnesium salt), glibenclamide and tolbutamide were purchased from Sigma. Ciclazindol was obtained from Wyeth, englitazone sodium was a gift from Pfizer and troglitazone a gift from Parke-Davis. Sodium azide was prepared as a 1 M stock solution in water, stock solutions of tolbutamide (100 mM), glibenclamide (10 mM), ciclazindol (10 mM) and englitazone (100 mM) were prepared in DMSO.

Electrophysiological recording and analysis

Oocytes:

Whole cell currents were measured using a two electrode voltage clamp technique. Voltages were applied and currents measured using a Macintosh computer running HEKA software connected to a GeneClamp 500 amplifier (Axon Instruments) via a Instutech ITC-16 interface. Both current and voltage electrodes were filled with 3 M KCl and had resistances of 0.5–2.0 MΩ. To minimize series resistance errors a virtual bath ground was used with a reference electrode in the bath and a second bath electrode, filled with 3 M KCl, positioned close to the oocyte. Recordings were made in a high-potassium bath solution, KD96 containing (mM): KCl 96, NaCl 2, CaCl₂ 1.8, HEPES 5 (pH 7.4 with KOH). Working concentrations of drugs were prepared in KD96 and superfused into the bath. Experiments were carried out at room temperature, 18–24°C. In order to construct concentration-response relations for drug-induced inhibition of whole-cell currents, the slope conductance was determined from the linear portion (−100 to −10 mV) of the current-voltage relation in control (with azide) and for a single concentration of drug. Concentration-response curves were fitted by non-linear regression to the following equation

where a=half maximal inhibitory concentration, b=drug concentration and n_H=Hill coefficient.

Single channel recordings were made from oocytes injected with Kir6.2 and SUR1 cRNA as above. Two to five days post-injection oocytes were placed in hypertonic stripping solution (which consisted of (mM): K⁺ aspartate 200, KCl 20, MgCl₂ 1, EGTA 10, HEPES 10 (pH 7.4 with KOH) for 5 min and the vitelline membrane removed manually. Naked oocytes were transferred to modified Barth's solution prior to use. Single channel recordings were made from excised patches in the inside-out configuration using electrodes with 5–8 MΩ resistances. Currents were recorded with a GeneClamp 500 amplifier (Axon Instruments), filtered at 1 kHz and recorded onto digital audio tape. The pipette solution contained (mM): KCl 140, MgCl₂ 1.2, CaCl₂ 2.6 and HEPES 10 (pH 7.4 with KOH) and the bath solution contained (mM): KCl 110, MgCl₂ 1.44, KOH 30, EGTA 10 and HEPES 10 (pH 7.2 with KOH).

Single channel recordings were made from inside-out patches obtained from transiently transfected HEK 293 cells selected as previously described. Recording pipettes (8–12 MΩ) were filled with solution containing (mM) KCl 140, CaCl₂ 1, MgCl₂ 1, HEPES-KOH 10 (pH 7.2), and the bath solution contained (mM) KCl 140, MgCl₂ 1, CaCl₂ 2, EGTA 10, HEPES-KOH 10. Single channel currents were measured using an EPC-7 patch clamp amplifier (List Electronics), filtered at 1 kHz and recorded onto digital audio tape, and replayed for illustration on a Gould TA240 chart recorder.

ATP and drugs were diluted from stock solutions into the internal solutions as indicated in the text and added to the bath by superfusion. Single channel data were analysed for current amplitude (I) and average channel activity (N_f.P_o, where N_f is the number of functional channels in the patch and P_o, the open state probability) as described previously (Lee et al., 1995). All data are expressed as the mean±s.e.mean and statistical analyses were performed using unpaired student's t-test.

Results

Oocytes injected with water, Kir6.2 or SUR1 alone or co-injected with Kir6.2 and SUR1 cRNAs exhibited no currents different to those of non-injected oocytes (Figure 1a). As previously reported (Gribble et al., 1997a) the application of sodium azide (3 mM) for 10–15 min induced significant current only from oocytes co-injected with both Kir6.2 and SUR1 (Figure 1b) and this was reversible on wash-out of the sodium azide (data not shown). The mean current amplitude at −100 mV was 0.027±0.03 μA before, and −12.9±0.9 μA (n=57) 15 min after application of 3 mM azide. This increase in current has been attributed to the azide causing metabolic inhibition, resulting in decreased intracellular ATP content, and subsequent activation of an ATP-sensitive K⁺ current (Gribble et al., 1997a). The whole-cell current induced by azide is almost completely blocked by application of 1 μM glibenclamide (n=3; Figure 1b) or 0.5 mM tolbutamide (n=6; Figure 1d), effects that are not readily reversible on washout of the drugs. Figure 1c shows the corresponding current (I)–voltage (V) relations for an oocyte injected with Kir6.2+SUR1 cRNA in the absence and presence of 3 mM sodium azide and following addition of 1 μM glibenclamide. Inhibition of these whole-cell currents by the sulphonylureas support the notion that azide has induced the activation of K_ATP channel currents in oocytes co-injected with Kir6.2 and SUR1.

Figure 1.

Activation of potassium current by sodium azide. (a,b) Whole cell currents recorded from two separate oocytes injected with (a) water and (b) cRNA encoding Kir6.2 and SUR1 before (control) and after exposure to 3 mM sodium azide for 15 min. Currents were elicited by a series of voltage steps from −120 mV to +30 mV from a holding potential of −10 mV, in this and subsequent figures. In (b) the azide-activated current is shown to be blocked following exposure to 1 μM glibenclamide in the presence of 3 mM sodium azide. (c) Corresponding current-voltage relations for the currents obtained in (b). (d) Concentration-inhibition curve for tolbutamide upon K_ATP channel currents activated by azide. Data are expressed as fractions of the control slope conductance (G_C). All points are the means of between four and six separate experiments and the vertical lines show the s.e.mean. The values for IC₅₀ were obtained by fitting the data by non-linear regression.

However, it has recently been demonstrated that activation of native K_ATP channels in the rat insulinoma cell line, CRI-G1, by sodium azide is not dependent upon a decline in cellular ATP content and that azide can activate K_ATP channel currents in isolated membrane patches (Harvey et al., 1999). Consequently we have examined the action of 3 mM sodium azide when applied directly to inside-out patches isolated from oocytes expressing Kir6.2-SUR1 currents. Inside-out patches excised from uninjected control oocytes exhibited no K_ATP channel activity (n=4). In contrast, patches excised from Kir6.2-SUR1 injected oocytes displayed significant channel activity, characterized, in symmetrical K⁺ recording conditions, by an approximately linear current-voltage relationship at hyperpolarized potentials in symmetrical K⁺ recording conditions with a mean single channel conductance of 69.1±0.01 pS (n=6) and exhibiting mild inward rectification at depolarized voltages (Figure 2a). In control experiments on inside-out patches, channel activity was demonstrated to be inhibited by ATP in a reversible manner with 0.5 and 1 mM ATP causing a 70.0±4.2% (n=3) and an 86.5±4.9% (n=3) inhibition of activity respectively (Figure 2b). Tolbutamide (0.5 mM) also inhibited these currents, causing a 94.9±1.4% (n=4) reduction in activity, an action reversible on washout of the tolbutamide in isolated patches. These data indicate that co-expression of Kir6.2 and SUR1 induces the appearance of single potassium channel currents with the properties particular of K_ATP channels. Application of 3 mM sodium azide, in the presence of 0.1 mM MgATP, to the cytoplasmic aspect of inside-out patches induced a rapid (between 30–120 s) increase (492±222%) in K_ATP channel activity (n=8; Figure 2c). The corresponding mean values of channel activity in the presence of 0.1 mM MgATP and following application of sodium azide were 0.50±0.24 and 1.94±0.72 respectively (n=8; P<0.05). The azide-induced increase in channel activity was readily reversed on washout of the azide, maintaining the presence of 0.1 mM MgATP (0.73±0.31; n=8). These data demonstrate that, in the presence of MgATP, sodium azide induces a rapid and reversible activation of Kir6.2-SUR1 currents in excised patches comparable to that reported for native K_ATP channels in CRI-G1 insulin-secreting cells (Harvey et al., 1999).

Figure 2.

Sodium azide activates single K_ATP channels. (a) Current-voltage relationship for K_ATP channel currents recorded, under symmetrical KCl conditions, from an inside-out patch excised from a Xenopus oocyte co-injected with cRNA encoding Kir6.2 and SUR1. (b) Continuous single channel currents recorded from an inside-out patch at a holding potential of −50 mV exposed to symmetrical 140 mM KCl. Application of either 1 mM Mg-ATP (top current trace) or 0.5 mM tolbutamide (lower current trace, separate patch) inhibited channel activity. The N_fP_o values were 2.59 and 0.30 in control and after addition of 1 mM Mg-ATP, which recovered to 3.41 on washout. The values were 1.41, 0.04 and 1.32 for control, 0.5 mM tolbutamide and wash respectively. (c) Single channel currents recorded from an inside-out patch under the same conditions as (b). >indicates continuity of the trace. Application of 0.1 mM Mg-ATP inhibited channel activity with N_fP_o decreasing from 3.36 to 0.20. Addition of 3 mM sodium azide in the continued presence of 0.1 mM Mg-ATP reversibly activated channel currents; N_fP_o increased to 4.07 in the presence of azide and decreased to 0.64 on wash out of sodium azide in the continued presence of 0.1 mM Mg-ATP. On complete wash out of the Mg-ATP, N_fP_o increased to 3.96.

Initial examination of Kir6.2-SUR1 pharmacology was performed on whole cell Kir6.2-SUR1 currents induced by 3 mM sodium azide. In the presence of azide, tolbutamide (0.1 μM–0.5 mM) was applied to voltage-clamped oocytes and engendered a concentration dependent, and irreversible, reduction in mean current (data not shown). From the tolbutamide concentration-inhibition curve, the IC₅₀ value and Hill coefficient were 15.3 μM and 0.84 respectively (Figure 1d) which are similar values to that reported for tolbutamide inhibition of native pancreatic β-cell K_ATP channels (Ashcroft & Ashcroft, 1990) and for Kir6.2-SUR1 currents expressed in Xenopus oocytes (Gribble et al., 1997c, 1998).

In marked contrast, Kir6.2-SUR1 currents are relatively insensitive to inhibition by ciclazindol in comparison to native K_ATP channels (Figure 3). For example, in a separate series of experiments on oocytes injected with Kir6.2-SUR1, the mean resting current, at a membrane potential of −100 mV, was 0.29±0.1 μA, which increased to 15.2±2.3 μA in the presence of 3 mM sodium azide. Application of 50 μM ciclazindol, a concentration sufficient to completely block native K_ATP channels in CRI-G1 insulin-secreting cells (Lee et al., 1996b), resulted in a mean whole cell current at −100 mV of 12.9±2.5 μA (n=8) a level not significantly (P>0.1) different from control. Higher concentrations of ciclazindol (0.5 mM) did significantly reduce the whole cell current, and from the concentration-response curve for azide-induced currents, the IC₅₀ value and Hill coefficient were 75.6 μM and 2.3 respectively (Figure 3c). This IC₅₀ value differs markedly from that reported previously using the CRI-G1 insulin-secreting cell line, values of 40 nM (Lee et al., 1996b) and 127 nM (Harvey & Ashford, 1998) being obtained.

Figure 3.

Effect of ciclazindol on whole cell currents. (a) Whole cell currents recorded from a Xenopus oocyte injected with cRNA encoding Kir6.2 and SUR1 under control conditions, after 15 min exposure to 3 mM sodium azide and 15 min later, following exposure to 50 μM ciclazindol in the presence of 3 mM sodium azide. (b) Corresponding current-voltage relations for the currents obtained in (a). (c) Concentration-inhibition curve for ciclazindol upon K_ATP channel currents activated by azide. Data are expressed as fractions of the control slope conductance (G_C). All points are the means of between three and seven separate experiments and the vertical lines show the s.e.mean. The values for IC₅₀ were obtained by fitting the data by non-linear regression. The dashed line represents the concentration-inhibition curve for ciclazindol on native K_ATP channels (from Lee et al., 1996b).

Similarly, the thiazolidinedione, englitazone at concentrations which reduce, significantly, native K_ATP channel currents had a much reduced inhibitory influence on Kir6.2-SUR1 currents activated by sodium azide in oocytes. For example, addition of 10 μM englitazone, a concentration close to that which half-maximally inhibits native K_ATP channel currents in insulin-secreting cells (Rowe et al., 1997; Harvey & Ashford, 1998), had no significant effect on azide-induced Kir6.2-SUR1 currents. However, application of higher concentrations of englitazone (50 and 100 μM) did inhibit Kir6.2-SUR1 whole cell currents (Figure 4a–c), an action not reversed on washout of drug, with an estimated IC₅₀ value and Hill coefficient of 93.6 μM and 1.7 respectively. Thus, azide-activated macroscopic Kir6.2-SUR1 currents expressed in oocytes do not completely replicate the pharmacological sensitivity observed for native pancreatic β-cell type K_ATP channels.

Figure 4.

Effect of englitazone on whole cell currents. (a) Whole cell currents recorded from a Xenopus oocyte injected with cRNA encoding Kir6.2 and SUR1 under control conditions, after 15 min exposure to 3 mM sodium azide and 15 min later, following exposure to 100 μM englitazone in the presence of 3 mM sodium azide. (b) Corresponding current-voltage relations for the currents obtained in (a). (c) Partial concentration-inhibition curve for englitazone upon K_ATP channel currents activated by azide. Data are expressed as fractions of the control slope conductance (G_C). All points are the means of six separate experiments and the vertical lines show the s.e.mean. The values for IC₅₀ were obtained by fitting the data by non-linear regression. The dashed line represents the concentration-inhibition curve for englitazone on native K_ATP channels (from Rowe et al., 1997).

It was possible that the presence of the sodium azide, used to activate Kir6.2-SUR1 currents in intact oocytes, was responsible for some or all of the difference in pharmacological sensitivity between the cloned channel currents and native K_ATP channels of pancreatic ß-cells. Consequently we examined the sensitivity of Kir6.2-SUR1 single channel currents in excised inside-out patches to inhibition by ciclazindol and englitazone in the absence and presence of sodium azide. Excised inside-out patch recordings were made in symmetrical (140 KCl in pipette and bath) K⁺-containing solution at a membrane potential of −50 mV. In the absence of sodium azide, application of 10 μM ciclazindol to Kir6.2-SUR1 currents resulted in a substantial inhibition (87.7±3.2%) of channel activity, an action reversible on washout of the drug (Figure 5a). The mean level of channel activity obtained in control conditions was 1.09±0.19, which declined to 0.12±0.03 in the presence of 10 μM ciclazindol (n=8; P<0.01). In contrast, in inside-out patches exposed to 3 mM sodium azide and 0.1 mM MgATP, addition of 10 μM ciclazindol resulted in no significant change in mean channel activity (Figure 5b); the mean value prior to ciclazindol was 1.66±0.54 and in the presence of 10 μM ciclazindol was 1.55±0.58 (n=12; P>0.8). A higher concentration of ciclazindol (100 μM) also had no significant effect on sodium azide-activated Kir6.2-SUR1 current (n=3; data not shown). Clearly, activation of the current by sodium azide dramatically depressed the efficacy of ciclazindol at inhibiting Kir6.2-SUR1 current. However this is not the case for englitazone as application of this drug to inside-out membrane patches inhibits Kir6.2-SUR1 currents in the absence and presence of sodium azide with equal potency. For example, 10 and 50 μM englitazone caused a 67.6±10.8% (n=3) and 81.5±5.8% (n=7) inhibition of Kir6.2-SUR1 currents in the absence of azide (Figure 5c) whereas in the presence of 3 mM sodium azide and 0.1 mM MgATP (Figure 5d) the inhibition was 47.0±10.3% (n=5) and 88.8±3.6% (n=5) respectively. This inhibition was also not reversed following washout of the drug.

Figure 5.

Sodium azide alters the sensitivity of single K_ATP channels to ciclazindol but not englitazone or tolbutamide. Single channel currents recorded from inside-out patches, excised from Xenopus oocytes co-injected with cRNA encoding Kir6.2 and SUR1, at a holding potential of −50 mV exposed to symmetrical 140 mM KCl. (a) In the absence of sodium azide, 10 μM ciclazindol (upper trace) inhibits cloned K_ATP channel activity. The N_fP_o values were 1.10, 0.11 and 0.90 for control, 10 μM ciclazindol and wash respectively. (b) Following activation by sodium azide in the presence of 0.1 mM Mg-ATP, 10 μM ciclazindol fails to inhibit cloned K_ATP channel currents. The N_fP_o values were 3.06, 2.71 and 2.24 for azide, 10 μM ciclazindol and wash respectively. In contrast the sensitivity to the sulphonylurea tolbutamide (lower trace) is unaffected by sodium azide activation. The N_fP_o values were 2.43, 0.41, and 2.55 for azide, 0.5 mM tolbutamide and wash respectively. (c) Englitazone inhibits cloned K_ATP channel activity in the absence of sodium azide. The N_fP_o values were 0.22 and 0.07 for control and on addition of 50 μM englitazone respectively, no wash out of the drug was observed. (d) Englitazone also inhibited cloned K_ATP channel activity following activation by sodium azide. The values of N_fP_o were 2.37 and 1.21 for azide and on addition of 10 μM englitazone respectively, again no wash out of the drug was observed.

These data indicate that the presence of sodium azide, used to activate Kir6.2-SUR1 currents in oocytes, results in a marked reduction in the efficacy of Kir6.2-SUR1 current to inhibition by ciclazindol,. The discrepancy in efficacy for englitazone between macroscopic currents recorded from intact oocytes and that of isolated patches from oocytes or K_ATP currents recorded from native tissue may be related to the use of the intact oocyte and/or slow access of the drug to its site of action. It has also been argued that both ciclazindol and englitazone inhibit pancreatic β-cell type K_ATP channels by acting at a site distinct from the sulphonylureas (Lee et al., 1996b; Rowe et al., 1997). Consequently, we have examined the actions of these drugs on cloned K_ATP channel currents, expressed in human embryonic kidney (HEK 293) cells, using the truncated form of Kir6.2 (Kir6.2ΔC26; the last 26 amino acids deleted from the C-terminus), expressed alone (Tucker et al., 1997) and co-expressed with SUR1, to determine which subunit confers drug sensitivity. Single channel currents were recorded, from inside-out membrane patches excised from HEK 293 cells transiently co-expressing Kir6.2ΔC26-SUR1 or Kir6.2ΔC26 alone, with the sensitivity of the currents to Mg-ATP being routinely examined to establish channel identity (Figures 6a and 7a). Application of 100 nM ciclazindol resulted in a significant (P<0.05) inhibition of channel activity when applied to patches expressing either Kir6.2ΔC26-SUR1 (Figure 6b) or Kir6.2ΔC26 currents (Figure 7b), inducing inhibition of channel activity by 79.0±3.1% (n=6) and 84.9±6.3% (n=4) respectively. Similarly, 10 μM englitazone induced significant (P<0.05) inhibition of both Kir6.2ΔC26-SUR1 (Figure 6c) and Kir6.2ΔC26 currents (Figure 7c); by 79.8±1.5% (n=5) and 93.8±2.5% (n=3) respectively. The effects of both these agents were not easily reversed following washout of the drugs. Therefore, ciclazindol and englitazone appear to be able to inhibit K_ATP currents expressed in HEK 293 cells independently of the presence of SUR1. Sodium azide (3 mM), in the presence of 0.1 mM Mg-ATP, did not increase channel currents composed of full length Kir6.2-SUR1 (n=3) or Kir6.2ΔC26-SUR1 (n=3) in inside-out patches in HEK 293 cells (data not shown).

Figure 6.

Single channel currents recorded from inside-out patches excised from HEK 293 cells transiently expressing Kir6.2Δ26+SUR1. Patches were held at a membrane potential of −40 mV in symmetrical 140 mM KCl. Application of (a) 0.1 mM Mg-ATP, (b) 100 nM ciclazindol and (c) 10 μM englitazone reduced channel activity, with poor recovery following washout of these agents. The values of N_fP_o were (a) 4.22 and 0.27 for control and 0.1 mM Mg-ATP respectively, (b) 4.69 and 0.37 for control and 100 nM ciclazindol respectively, and (c) 6.84 and 1.53 for control and 10 μM englitazone respectively.

Figure 7.

Single channel currents recorded from inside-out patches excised from HEK 293 cells transiently expressing Kir6.2Δ26 alone. Patches were held at a membrane potential of −40 mV in symmetrical 140 mM KCl. Application of (a) 0.1 mM Mg-ATP, (b) 100 nM ciclazindol and (c) 10 μM englitazone reduced channel activity, with poor recovery following washout of these agents. The values of N_fP_o were (a) 0.96 and 0.04 for control and 0.1 mM Mg-ATP respectively, (b) 0.45 and 0.004 for control and 100 nM ciclazindol respectively, and (c) 0.34 and 0.01 for control and 10 μM englitazone respectively.

Discussion

The characteristics of the currents induced by co-expression of cRNAs encoding Kir6.2 and SUR1 in Xenopus oocytes are in agreement with the previously published reports (Gribble et al., 1997a,1997b). The sensitivity of the expressed channel to agents that act to inhibit channel activity is similar to that of the native K_ATP channels of pancreatic β cells with respect to MgATP and the sulphonylureas glibenclamide and tolbutamide. However, the actions of sodium azide and the resultant alterations in the apparent pharmacological specificity and sensitivity of Kir6.2-SUR1 macroscopic currents raises some doubts as to the suitability of this method of evaluating drug potency. Clearly, the inhibitory actions of ciclazindol and englitazone, two structurally unrelated inhibitors of native pancreatic β-cell type K_ATP channels, are substantially altered when examined by this method.

Current convention propounds that cloned K_ATP whole-cell currents in oocytes are inhibited at resting intracellular ATP concentrations and can only be observed following metabolic inhibition of the oocyte (Gribble et al., 1997a). The metabolic inhibitor sodium azide, which has been routinely used to activate cloned K_ATP whole-cell currents, is proposed to inhibit mitochondrial electron transport resulting in a lowering of intracellular ATP levels. We demonstrate, however, that sodium azide not only can activate whole cell Kir6.2-SUR1 currents but can also, in the presence of MgATP, induce a rapid, reversible activation of Kir6.2-SUR1 currents in excised patches. This latter effect is not expected following the metabolic inhibition argument. Furthermore, levels of total ATP concentration in oocytes are not dramatically reduced by metabolic inhibition, even for periods up to 60 min (Gribble et al., 1997a). In a recent study it was also demonstrated that, in the presence of MgATP, sodium azide activates K_ATP channels in excised patches from CRI-G1 insulin-secreting cells (Harvey et al., 1999). Interestingly, this effect may not be isolated to sodium azide as 2,4 dinitrophenol (DNP) has been demonstrated to activate cardiac K_ATP channels independently of metabolic inhibition (Alekseev et al., 1997). Thus azide, like DNP, may activate K_ATP channels by an alternative mechanism unrelated to intracellular ATP depletion. Surprisingly, however, sodium azide had no activating action on isolated patches containing Kir6.2-SUR1 currents (or indeed on the truncated Kir6.2ΔC26 currents) when transiently expressed in HEK cells, suggesting that co-expression of Kir6.2 and SUR1 in HEK 293 may not fully reconstitute the native environment for these channels. One explanation for these results is that another protein or factor is required in order to observe the sodium azide activation and this factor is present in the CRI-G1 insulin-secreting cells and Xenopus oocytes, but not the HEK cells used in the present study, or that it is lost/inactivated in the inside-out configuration.

The sensitivity of both pancreatic β-cell type K_ATP channel currents and heterologously expressed Kir6.2-SUR1 currents to tolbutamide was unaltered following azide activation. This is in contrast to the action of DNP which is reported to occlude sulphonylurea inhibition of K_ATP channel currents (Findlay, 1993; Guillemare et al., 1995; Alekseev et al., 1997). In contrast, the macroscopic Kir6.2-SUR1 currents, recorded from oocytes, are relatively insensitive to block by both ciclazindol and englitazone, compounds which inhibit, significantly, native K_ATP channels at the concentrations used in this series of experiments. It is clear from the inside-out patch experiments that the presence of azide markedly affects the sensitivity of the cloned channels to ciclazindol and it is likely that the same effect accounts for the reduced efficacy of this drug on the whole cell currents. The same is not true of englitazone where, although the whole cell currents show reduced sensitivity to this compound, inside-out patches exhibit sensitivity close to that of native K_ATP channels. It is possible that this apparent conflict is due to poor access of the drug in the intact oocyte.

Interestingly, the ob gene product, leptin, which activates K_ATP channels in both hypothalamic glucose-responsive neurones (Spanswick et al., 1997) and CRI-G1 cells (Harvey et al., 1997) has also been reported to cause a substantial reduction in the sensitivity of CRI-G1 K_ATP channel currents to these inhibitors (Harvey & Ashford, 1998). This latter study, which compared the effects of leptin and diazoxide, demonstrated that ciclazindol sensitivity was dramatically reduced following leptin application such that the EC₅₀ values were 127 nM and 40 μM for diazoxide activated and leptin activated K_ATP channels respectively. Leptin also altered the sensitivity to englitazone, with EC₅₀ values of 7.7 μM and 52 μM for diazoxide- and leptin-activated K_ATP channels respectively. Although we have previously alluded to the sodium azide-induced K_ATP channel activation having a great deal in common with that of diazoxide activation (Harvey et al., 1999), the results obtained above, particularly with respect to inhibition of K_ATP channel currents by ciclazindol, indicate that there may also be a common process in the mechanism by which azide and leptin activate K_ATP channels.

Sulphonylureas have been clearly shown to act via the high affinity sulphonylurea receptor, SUR1, but the sites of action by which ciclazindol and englitazone mediate their effects on K_ATP channels are unclear, although there is evidence that these sites are distinct from that of the sulphonylureas. For example previous studies indicate that treatments which cause functional uncoupling of K_ATP channel currents from inhibition by the sulphonylureas do not alter the channel sensitivity to inhibition by ciclazindol or englitazone (Lee et al., 1996b, Rowe et al., 1997). In addition ciclazindol does not displace ³H-glibencamide binding from CRI-G1 (Lee et al., 1996b) or porcine brain (Noack et al., 1992) membranes. The present results using HEK 293 cells transfected with a truncated form of Kir6.2, Kir6.2ΔC26, demonstrate unequivocally that the site of action of both ciclazindol and englitazone is distinct from the sulphonylurea receptor and most likely lies on the Kir6.2 subunit. Furthermore, the presence of SUR1 appears to have no action on the potency of these drugs to inhibit currents generated by Kir6.2. This mirrors the results obtained for cibenzoline inhibition of cloned K_ATP channel currents where it was demonstrated to be equally effective at inhibiting Kir6.2ΔC26 and Kir6.2ΔC26+ SUR1 currents (Mukai et al., 1998). However, the observation that sodium azide-activated Kir6.2-SUR1 currents expressed in Xenopus oocytes are differentially sensitive to ciclazindol, but not englitazone, in the isolated patch configuration indicates that these inhibitors are unlikely to be acting at the same site on Kir6.2.

Abbreviations

HEK

human embryonic kidney

K_ATP

ATP-sensitive potassium channel

Kir6.2

inward rectifier potassium channel subunit 6.2

SUR1

sulphonylurea receptor type 1