Differential Interactions of Nateglinide and Repaglinide on the Human -Cell Sulphonylurea Receptor 1

Home , Glibenclamide, Mitiglinide, Nateglinide, Repaglinide, Tolbutamide

Differential Interactions of Nateglinide and Repaglinide on the Human ␤-Cell Sulphonylurea Receptor 1 Ann Maria K. Hansen,1 Inge T. Christensen,1 John Bondo Hansen,1 Richard D. Carr,1 Frances M. Ashcroft,2 and Philip Wahl1

Repaglinide and nateglinide represent a new class of plasma glucose levels potentiates GLUT2-mediated glu- insulin secretagogues, structurally unrelated to sulpho- cose uptake into the ␤-cell, and subsequent metabolism of nylureas, that were developed for the treatment of type glucose increases the ATP/ADP ratio, leading to closure of 2 diabetes. The inhibitory effect of these drugs was KATP channels, membrane depolarization, and opening of investigated on recombinant wild-type and mutant voltage-sensitive Ca2ϩ channels. The resulting increase in Kir6.2/SUR1 channels expressed in HEK293 cells. intracellular Ca2ϩ stimulates insulin secretion. Sulphonyl- Nateglinide and repaglinide dose-dependently inhibited ureas trigger the same series of events by blocking K whole-cell Kir6.2/SUR1 currents with half-maximal in- ATP channels directly. KATP channels are also found in a variety hibitory concentration (IC50) values of 800 and 21 nmol/l, respectively. Mutation of serine 1237 in SUR1 to of other tissues, including neurons, heart, and skeletal and tyrosine (S1237Y) abolished tolbutamide and nateglin- smooth muscle cells, where they play important physio- ide block, suggesting that these drugs share a common logical and pathophysiological roles (4). point of interaction on the SUR1 subunit of the ATP- The ␤-cell K channel is a hetero-octameric complex ؉ ATP sensitive K channel. In contrast, repaglinide inhibition of four inwardly rectifying potassium channel (Kir6.2) ؍ was unaffected by the S1237Y mutation (IC50 23 subunits, which form a tetrameric pore, and four regula- nmol/l). Radioligand binding studies revealed a single tory sulphonylurea receptor (SUR1) subunits (5,6). Sul- high-affinity binding site for [3H]repaglinide on membranes prepared from HEK293 cells expressing wild- phonylureas bind with high affinity to SUR1 to mediate closure of the Kir6.2 pore (7). Different SUR variants ؍ type (equilibrium dissociation constant [KD] 0.40 ؍ nmol/l) or mutant (KD 0.31 nmol/l) Kir6.2/SUR1 chan- confer different sensitivities to sulphonylureas and KATP nels. Nateglinide and tolbutamide displaced [3H]repa- channel openers on the Kir6.2 subunit; for example, tolbu- glinide binding to wild-type channels with IC50 values of tamide inhibits SUR1- but not SUR2-containing KATP chan- 0.7 and 26 ␮mol/l, respectively, but produced <10% nels (8,9). This accounts for the much lower potency of 3 displacement of [ H]repaglinide bound to mutant chan- this drug in tissues such as heart and smooth muscles, in nels. This is consistent with the idea that binding of which K channels are composed of SUR2A and SUR2B, nateglinide and tolbutamide, but not repaglinide, is ATP abolished by the SUR1[S1237Y] mutation and that the respectively (10,11). binding site for repaglinide is not identical to that of In addition to the classic sulphonylureas, KATP channels nateglinde/tolbutamide. These results are discussed in are inhibited by benzamido compounds and their deriva- terms of a conformational analysis of the drug mole- tives (e.g., meglitinide) (12). This suggests that like other cules. Diabetes 51:2789–2795, 2002 ATP-binding cassette transporters, SUR possesses a large multifaceted drug-binding pocket that can accommodate several structurally distinct compounds. Individual compounds differ in the residues within this pocket, with ulphonylureas such as tolbutamide and gliben- which they interact. Studies of recombinant K channels clamide are widely used to treat type 2 diabetes ATP suggest that drugs containing a sulphonylurea moiety (e.g., because they stimulate insulin secretion. Their primary mode of action is to bind to ATP-sensi- tolbutamide, glibenclamide) interact with residues in the Sϩ TM15-16 linker of SUR1 and that a single serine residue tive K (K ) channels in the pancreatic ␤-cell and induce ATP (S1237) within this region is critical for drug binding their closure (1–4). KATP channels serve a critical role in glucose-stimulated insulin secretion. At low glucose levels, and/or transduction (8,13). Mutagenesis and affinity-label- these channels are open, permitting an efflux of Kϩ ions ing studies also suggest that residues within the cytosolic that hyperpolarizes the ␤-cell membrane. Elevation of loop linking transmembranes (TMs) 5 and 6 may be essential for [3H]glibenclamide binding (14,15). Thus, amino acid residues derived from two distinct regions of From 1Discovery, Novo Nordisk A/S, Bagsvaerd, Denmark; and the 2University SUR1 appear to contribute to the glibenclamide binding Laboratory of Physiology, Oxford University, Oxford, U.K. Address correspondence and reprint requests to Dr. Philip Wahl, Depart- site. Because glibenclamide contains both a sulphonylurea ment of Islet Discovery Research, Discovery, Novo Alle´, Novo Nordisk A/S, group and a benzamido moiety, these moieties may inter- DK-2880 Bagsvaerd, Denmark. E-mail: [email protected]. Received for publication 21 December 2001 and accepted in revised form 7 act with different parts of SUR1, with the sulphonylurea June 2002. moiety interacting with the TM15-16 linker. F.M.A. is a paid consultant for Novo Nordisk. The extent to which this model can be applied to drugs GFP, green fluorescent protein; IC50, half-maximal inhibitory concentration; ϩ that do not contain a sulphonylurea group remains un- KATP, ATP-sensitive K ;KD, equilibrium dissociation constant; SUR, sulphonylurea receptor; TM, transmembrane. clear. Recent studies have shown that inhibition of Kir6.2/

DIABETES, VOL. 51, SEPTEMBER 2002 2789 INTERACTION OF REPAGLINIDE AND NATEGLINIDE WITH SUR1

SUR1 currents by the nonsulphonylurea mitiglinide, as time course of drug inhibition. In this case, currents were sampled at 25 Hz with tolbutamide, is abolished by the S1237Y mutation and filtered at 6.25 Hz. Molecular modeling. Conformational analyses were performed using Macro- (16). This suggests that this residue is not specific for the Model version 6.0 (20). The compounds were subjected to a Monte Carlo sulphonylurea moiety but can interact with other com- search using the MMFF force field (21), with a solvation model for water (22). pounds. We therefore explored whether this part of the For each molecule, 5,000 conformations were generated, with an energy drug-binding pocket of SUR1 can also accommodate repa- cut-off of 25 kJ/mol. The resulting low-energy conformations were imported into Sybyl version 6.6 (Tripos, St. Louis, MO), and all further manipulations glinide and nateglinide, two members of a new class of were performed using this program. Initially, a number of alternative super- insulin secretagogues termed “prandial glucose regula- impositions were generated using the GASP algorithm (23), as implemented in tors.” These drugs differ structurally from the sulphonyl- Sybyl. The final pharmacophore models were generated by manually adjusting ureas in that they do not possess a sulphonylurea moiety. the superimpositions generated using GASP. In all cases, the resulting Chemically, they are based on benzoic acid (in the case of conformations were compared with the minimum-energy conformations found in the conformational search. repaglinide [17]) and phenylalanine (nateglinide). Both Data analysis. Data are presented as mean Ϯ 1 SD unless otherwise stated.

drugs are known to inhibit KATP channel activity. The Concentration-response curves for drug-induced KATP current inhibition were half-maximal inhibitory concentration (IC50) for nateglin- constructed by expressing the current in the presence of the drug as a fraction ␤ ␮ of the current before the drug was added. Data were analyzed in Prism ide inhibition of native rat -cell KATP channels (7 mol/l) (18) lies, as with tolbutamide, in the low micromolar (GraphPad) using the four-parameter logistic equation: range. In contrast, repaglinide resembles glibenclamide in 100 Ϫ a y ϭ a ϩ ͑ Ϫ͓ ͔ ⅐ producing high-affinity block of both native and recombi- 1 ϩ 10 log IC50 L nH ␤ ϭ nant -cell KATP channels (IC50 0.9–7 nmol/l) (18,19). Our results suggest that nateglinide, but not repaglinide, where y is the current expressed as a percentage of that recorded before drug was added, a is the percentage of current remaining after maximal inhibition interacts with serine 1237 of SUR1 to mediate inhibition of by the drug, IC50 is the drug concentration that results in half-maximal the KATP channel. inhibition, [L] is the concentration of drug, and nH is the Hill coefficient. The same equation was used to fit the dose-inhibition curves for displacement of [3H]repaglinide binding, but in this case, y was specific binding and a RESEARCH DESIGN AND METHODS was nonspecific binding. The experimentally measured IC50 values for the Molecular biology. Human SUR1 cDNA (GenBank L78207) and human Kir6.2 competitive ligands were converted into Ki values using the Cheng-Prusoff (GenBank D50582) were cloned into pcDNA3.1(Ϫ) (Invitrogen). The point equation: mutation SUR1[S1237Y] was constructed by standard molecular biology techniques and confirmed by DNA sequencing. IC ϭ 50 Cell culture and transfection. HEK293 cells were cultured at 37°Cina Ki 1 ϩ ͓L͔/KD humidified atmosphere of 95% air and 5% CO2 in Dulbecco’s modified Eagles 3 medium with 4.5 g/l glucose (BioWhittaker) supplemented with 10% FCS, where [L] is the concentration of [ H]repaglinide, and KD is the equilibrium penicillin (100 units/ml), and streptomycin (0.1 mg/ml). dissociation constant for [3H]repaglinide. Transient transfections were performed using FuGene 6 Transfection Chemicals. Tolbutamide was purchased from Sigma and glibenclamide from Reagent (Roche), according to the manufacturer’s instruction. Cells were Research Biochemicals International. Nateglinide was synthesized at Novo seeded at 50% confluency and transfected with Kir6.2 and SUR1[S1237Y] at a Nordisk A/S and repaglinide at Boehringer Ingelheim (Biberach, Riss, Germa- plasmid ratio of 1:3 on the next day. Cells to be used for electrophysiological ny). Concentrated stock solutions were prepared in DMSO for subsequent experiments were also cotransfected with green fluorescent protein (GFP) to dilution in buffer. The concentration of DMSO in the experiments did not enable visual identification of transfected cells. Experiments were performed exceed 0.1% and had no effect in either binding or electrophysiological 1–3 days after transfection. HEK293 cells stably expressing hKir6.2 and hSUR1 experiments. were used for studies of wild-type channels (19). Radiolabelled repaglinide was prepared at the Department of Isotope Membrane preparation. Cells were harvested and centrifuged at 48,000g for Chemistry, Novo Nordisk A/S, by catalytic tritiation of the repaglinide 10 min at 4°C. The pellet was homogenized in ice-cold buffer (30 mmol/l precursor S(ϩ)-2-ethoxy-4-[2-[[3-methyl-1-[2-(piperidinyl)-phenyl]4-buten-yl] Tris-HCl, pH 7.4) using an Ultra Turrax for 20 s. Centrifugation and homoge- amino]2-oxoethyl]-benzoic acid (17), which was kindly provided by Dr. M. nization were repeated, and the pellet was then resuspended in buffer and Mark, Boehringer Ingelheim. A specific activity of ϳ12 MBq/␮g (144 Ci/mmol) sucrose was added to a final concentration of 250 mmol/l. Protein concentra- was estimated from mass spectroscopy of the final product. tion was measured using the Bio-Rad protein assay. Membranes were stored at Ϫ80°C. RESULTS Binding experiments with [3H]repaglinide. Binding experiments were performed in 96-well OptiPlates (Packard). Membranes (5 ␮g protein/well for Electrophysiology. Whole-cell currents were recorded wild-type channels and 12 ␮g protein/well for mutant channels) were thawed from HEK293 cells coexpressing Kir6.2 and either SUR1 or 3 on ice and incubated for 60 min at 37°C with 0.8 nmol/l [ H]repaglinide in 30 SUR1[S1237Y]. After establishment of the whole-cell con- mmol/l HEPES (pH 7.4), in a total volume of 250 ␮l. Bound [3H]repaglinide was separated from free [3H]repaglinide by rapid filtration on a Filtermate figuration and dialysis with intracellular solution, there Harvester (Packard) through UniFilter GF/B filterplates (Packard). Filter- was a gradual increase in both Kir6.2/SUR1 and Kir6.2/ ␮ plates were washed five times with 450 l water (room temperature) and SUR1[S1237Y] currents due to opening of KATP channels. dried. Scintillation cocktail (30 ␮l) (Microscint, Packard) was added to each As shown in Fig. 1A, tolbutamide (300 ␮mol/l) completely well, and radioactivity was determined by counting the plates in a Microplate Ϯ ϭ Scintillation and Luminescence Counter (Topcount-NTX; Packard). Nonspe- inhibited this KATP current (by 93 4%, n 4), and the cific binding was determined in the presence of 1 ␮mol/l unlabeled repaglinide effect of the drug was reversible upon return to the control and was Ͻ5% of the total binding. Binding experiments were performed in solution. In contrast, tolbutamide had very little effect on triplicate (Kir6.2/SUR1) or duplicate (Kir6.2/SUR1[S1237Y]). Kir6.2/SUR1[S1237Y] currents (Fig. 1B). Glibenclamide (1 Electrophysiology. Whole-cell currents were recorded at 20–22°C using an ␮mol/l) blocked Kir6.2/SUR1 channels by 98 Ϯ 1% (n ϭ 3), EPC9 patch-clamp amplifier and PulseϩPulseFit version 8.07 software. The extracellular bath solution contained 140 mmol/l NaCl, 5 mmol/l KCl, 10 and this inhibition could not be reversed by 20 min of

mmol/l HEPES, 1.8 mmol/l CaCl2, and 20 mmol/l mannitol (pH 7.4 with NaOH). washing (Fig. 1A). Although glibenclamide also inhibited Cells were dialysed with intracellular solution containing 120 mmol/l KCl, 1 Kir6.2/SUR1[S1237Y] channels (by 67 Ϯ 6%; n ϭ 3), in this mmol/l MgCl , 5 mmol/l EGTA, 2 mmol/l CaCl , 20 mmol/l HEPES (pH 7.3 with 2 2 case, inhibition was largely reversed on wash-out of the Ϫ KOH), 0.3 mmol/l K2 ATP, and 0.3 mmol/l K2 ADP. Cells were clamped at 80 mV, and currents were evoked by repetitive 200 ms, 10 mV depolarizing drug (Fig. 1B). These ﬁndings are in agreement with earlier voltage steps. Signals were sampled at 20 kHz and ﬁltered at 5 kHz. In some studies of rodent Kir6.2/SUR1 channels expressed in Xe- experiments, cells were held at Ϫ70 mV for a period of 10–20 min to study the nopus oocytes (9).

2790 DIABETES, VOL. 51, SEPTEMBER 2002 A.M.K. HANSEN AND ASSOCIATES

FIG. 1. Representative whole-cell recordings of the inhibition of wild- ␮ type and mutant KATP currents by tolbutamide (300 mol/l) or glibenclamide (1 ␮mol/l). HEK 293 cells expressing Kir6.2/SUR1 (A)or Kir6.2/SUR1[S1237Y] (B) channels were clamped at ؊70 mV. The horizontal bar indicates 100 s and the vertical bar 200 pA. The dotted line indicates the zero current level and the thick horizontal bar indicates the time of application of the drug.

Figure 2 compares the effects of repaglinide and nateglinide on wild-type and mutant KATP channels. Repaglinide (1 ␮mol/l) completely inhibited human Kir6.2/SUR1 currents (by 98 Ϯ 1%, n ϭ 5) (Fig. 2A), as previously reported for rat Kir6.2/SUR1 expressed in Xenopus oocytes (19). A similar extent of block was observed for Kir6.2/ SUR1[S1237Y] (94 Ϯ 1%, n ϭ 4) (Fig. 2B), suggesting that FIG. 3. Concentration-response curves for inhibition of Kir6.2/SUR1 S1237 is not required for repaglinide inhibition. In both (f/F) and Kir6.2/SUR1[S1237Y] (Ⅺ/E) channels by repaglinide (f/Ⅺ) or nateglinide (F/E). The current in the presence of drug (I) is cases, no significant recovery of current was observed on expressed as a percentage of that recorded before the drug was added removal of the drug. In contrast, nateglinide (100 ␮mol/l) (Icontrol). Data points represent the means ؎ SEM of three to eight produced reversible inhibition of Kir6.2/SUR1 currents experiments. (96 Ϯ 2%, n ϭ 4) (Fig. 2A) but was without significant effect on Kir6.2/SUR1[S1237Y] channels (Fig. 2B). (18–28), respectively. These results suggest that S1237 is Concentration-response relationships for repaglinide not critical for repaglinide-induced channel inhibition. and nateglinide block of wild-type and mutant K chan- In contrast to repaglinide, nateglinide blocked Kir6.2/ ATP ␮ nels are shown in Fig. 3. Repaglinide blocked Kir6.2/SUR1 SUR1 currents with an IC50 of 0.8 mol/l (95% CI 0.3–2.3) and Kir6.2/SUR1[S1237Y] currents with similar potency: but produced Ͻ10% block of the mutant channel, even at ␮ IC ϭ 21 nmol/l (95% CI 17–26) and IC ϭ 23 nmol/l a concentration of 100 mol/l. Similar IC50 values were 50 50 found for Kir6.2/SUR1 expressed in COS-1 cells (24) and ␤ for native rat -cell KATP currents (25). This suggests that the presence of a tyrosine at residue 1237 in SUR1 either prevents nateglinide binding or the transduction of binding into channel closure. [3H]Repaglinide binding experiments. High levels of specific[3H]repaglinide binding were observed for membranes prepared from HEK293 cells expressing Kir6.2/ SUR1 but not for membranes isolated from HEK293 cells expressing the vector only. Saturation binding experiments showed that [3H]repaglinide binding was saturable, and a Scatchard transformation of the data indicated that binding to Kir6.2/SUR1 occurred in an apparent monopha- ϭ Ϯ sic manner with a high capacity (Bmax 5.1 1.8 pmol/mg ϭ protein, n 5) (Fig. 4A). The KD, estimated from nonlinear regression, was 0.40 Ϯ 0.09 nmol/l (n ϭ 5). Similarly, binding of [3H]repaglinide to Kir6.2/SUR1[S1237Y] re- FIG. 2. Representative whole-cell recordings of the inhibition of wild- Ϯ type and mutant K currents by repaglinide (1 ␮mol/l) and nateglin- vealed a single binding site (Fig. 4B)withaKD of 0.31 ATP Ϯ ϭ ide (1 ␮mol/l). HEK 293 cells expressing Kir6.2/SUR1 (A) or Kir6.2/ 0.02 nmol/l and a Bmax of 1.6 0.2 pmol/mg protein (n SUR1[S1237Y] (B) channels were clamped at ؊70 mV. The horizontal 3). Thus, repaglinide affinity for Kir6.2/SUR1 is unaffected bar indicates 100 s and the vertical bar 200 pA. The dotted line indicates the zero current level and the thick horizontal bar indicates by mutation of S1237 to tyrosine. The difference observed the time of application of the drug. in Bmax for wild-type and mutant channels is likely to be

DIABETES, VOL. 51, SEPTEMBER 2002 2791 INTERACTION OF REPAGLINIDE AND NATEGLINIDE WITH SUR1

FIG. 5. Competition binding to membranes from HEK 293 cells expressing Kir6.2/SUR1 (A) or Kir6.2/SUR1[S1237Y] (B). Displacement of specific[3H]repaglinide with repaglinide (f), glibenclamide (Œ), tol- FIG. 4. Saturation binding of [3H]repaglinide to membranes prepared butamide (᭜), or nateglinide (F). Results are expressed as percentage from HEK 293 cells expressing Kir6.2/SUR1 (A) or Kir6.2/ of specific binding in the absence of competing drug. Data are from a SUR1[S1237Y] (B). Data are from a single representative experiment single representative experiment in which data points were collected in which data points were collected in triplicate (Kir6.2/SUR1) or in triplicate (Kir6.2/SUR1) or duplicate (Kir62/SUR1[S1237Y]). duplicate (Kir6.2/SUR1[S1237Y]). order of potency for these drugs on inhibition of Kir6.2/ the result of the different expression systems (transient SUR1 currents (26). versus stable transfection). In the case of Kir6.2/SUR1[S1237Y], repaglinide dis- 3 3 Ϯ Attempts to measure [ H]nateglinide binding to wild- placed [ H]repaglinide binding with a Ki of 0.4 0.2 nmol/l type or mutant Kir6.2/SUR1 were not successful, possibly (n ϭ 3), which is similar to that found for the wild-type ϭ because of the low affinity and rapid unbinding of the drug channel. In contrast, the affinity for glibenclamide (Ki (18). We therefore examined the ability of unlabelled 36 Ϯ 6 nmol/l, n ϭ 3) was 170-fold lower than that of the nateglinide to displace [3H]repaglinide binding to mem- wild-type channel. Furthermore, nateglinide did not dis- branes isolated from HEK293 cells expressing Kir6.2/SUR1 place [3H]repaglinide binding to the mutant channel, even or Kir6.2/SUR1[S1237Y]. For comparative purposes, we at the highest concentration tested (30 ␮mol/l), and tolbu- also examined the effects of cold repaglinide, gliben- tamide (300 ␮mol/l) only marginally displaced [3H]repa- clamide, and tolbutamide on [3H]repaglinide binding. glinide binding. The mean results are summarized in Table All drugs showed a monophasic displacement of 1. The data are consistent with the idea that the nateglinide 3 [ H]repaglinide binding (Fig. 5). The Ki for the wild-type binding, as with that of tolbutamide, is abolished by the channel, estimated from the IC50 values (see RESEARCH S1237Y mutation in SUR1. Furthermore, this mutation DESIGN AND METHODS), were 0.6 Ϯ 0.3 nmol/l (n ϭ 3) for reduces the affinity for glibenclamide binding while leav- repaglinide, 0.2 Ϯ 0.1 nmol/l (n ϭ 3) for glibenclamide, ing that of repaglinide unaffected. 0.24 Ϯ 0.04 ␮mol/l (n ϭ 3) for nateglinide, and 9.0 Ϯ 3.2 Structural comparison. The fact that S1237 is critical for ␮mol/l (n ϭ 3) for tolbutamide. The Hill coefficients were nateglinide as well as tolbutamide inhibition raises the close to unity in all cases (from 0.90 to 1.19). The Ki for question of whether these drugs show structural similari- repaglinide (0.6 Ϯ 0.3 nmol/l) is not significantly different ties. We therefore carried out structural comparisons of from the KD estimated from the saturation experiments repaglinide, glibenclamide, nateglinide, and tolbutamide (0.40 Ϯ 0.09 nmol/l). Furthermore, the relative order of (Fig. 6A). affinity is glibenclamide ϳ repaglinide Ͼ nateglinide Ͼ In agreement with earlier studies (17), repaglinide and tolbutamide, and is in close agreement with the relative glibenclamide can be fitted to the same pharmacophore

2792 DIABETES, VOL. 51, SEPTEMBER 2002 A.M.K. HANSEN AND ASSOCIATES

TABLE 1 Comparison of [3H]repaglinide binding data on Kir6.2/SUR1 and Kir6.2/SUR1[S1237Y] Kir6.2/SUR1 Kir6.2/SUR1[S1237Y]

Compound IC50 (nmol/l) nH Ki (nmol/l) IC50 (nmol/l) nH Ki (nmol/l) Repaglinide 1.9 Ϯ 0.8 Ϫ1.19 Ϯ 0.10 0.6 Ϯ 0.3 1.2 Ϯ 0.7 Ϫ1.05 Ϯ 0.14 0.4 Ϯ 0.2 Glibenclamide 0.7 Ϯ 0.2 Ϫ1.14 Ϯ 0.13 0.2 Ϯ 0.1 105 Ϯ 17 Ϫ0.90 Ϯ 0.25 36 Ϯ 6 Nateglinide 679 Ϯ 121 Ϫ0.98 Ϯ 0.05 235 Ϯ 42 Ͼ30,000 N/A N/A Tolbutamide 26,000 Ϯ 9,300 Ϫ0.91 Ϯ 0.12 9,000 Ϯ 3,220 Ͼ300,000 N/A N/A Ϯ ϭ Data are means SD (n 3). N/A, not applicable; nH, Hill coefficient. model, with the benzoic acid fragment of repaglinide tolbutamide and nateglinide are smaller than glibenclamide, superpositioned on the acidic arylsulphonamide part of the benzamide part of glibenclamide is depicted in a random glibenclamide (Fig. 6B). The distal aromatic rings are extended conformation. copositioned, whereas the carbonyl oxygens of the amide groups are capable of interacting with the same putative DISCUSSION hydrogen bond donor in SUR1. As both molecules are very flexible, the model shows only one of the several alterna- The ability of the new insulin secretagogues repaglinide tive spatial arrangements of the three pharmacophore and nateglinide to inhibit recombinant human Kir6.2/SUR1 groups (benzoic acid/arylsulphonamide, distal aromatic channels was investigated and compared with that of ring, and carbonyl oxygen). tolbutamide and glibenclamide. Structurally, gliben- Figure 6C shows that although nateglinide does not con- clamide consists of a sulphonylurea group similar to tain a sulphonylurea moiety, its three-dimensional structure tolbutamide and a nonsulphonylurea moiety that resem- is similar to that of tolbutamide and to the sulphonylurea part bles the benzoic acid derivative meglitinide. Previous of glibenclamide. In this model, the phenylpropionic acid mutagenesis and affinity-labeling studies have provided part of nateglinide is superimposed on the arylsulphonamide evidence for a model of the glibenclamide binding site in part of the sulphonylureas. In addition, the carbonyl oxygens which contributing amino acid residues are derived from and the hydrophobic tails are superpositioned. As both two distinct regions of SUR1: the cytosolic loop linking TMs 15 and 16 (in which S1237 is critical) and the cytosolic loop linking TMs 5 and 6 (8,14,15). The TM15-16 loop is essential for high-affinity inhibition of KATP channel currents by drugs that resemble tolbutamide in structure. In the present study, three different lines of evidence suggest that nateglinide, but not repaglinide, interacts with this region of the channel: binding data, electrophysiological data, and structural considerations. 3 Binding data. The KD for [ H]repaglinide binding ob- tained in this study (0.4 nmol/l) is lower than that reported for [3H]repaglinide binding to intact ␤TC3 cells (6.4 nmol/l) (27) but similar to the Ki (1.8 nmol/l) found for repaglinide inhibition of [3H]glibenclamide binding to RIN- 3 m5F cells (18). The Ki for inhibition of [ H]repaglinide binding to human Kir6.2/SUR1 by nateglinide presented here (235 nmol/l) is in accordance with that of human ϭ SUR1 expressed in COS-1 cells (IC50 280 nmol/l) (24) ϭ and that of native KATP channels in HIT-T15 cells (Ki 248 ϭ nmol/l [28] or 434 nmol/l [29]) or RIN-m5F cells (Ki 170 nmol/l [18]). In the present study, tolbutamide and nateglinide were found to fully displace [3H]repaglinide binding to Kir6.2/SUR1 channels. This suggests that the binding sites for tolbutamide/nateglinide and repaglinide either overlap or are allosterically coupled. Electrophysiology. Mutation of serine 1237 in SUR1 to tyrosine, which is known to influence channel inhibition by tolbutamide and glibenclamide (8), abolished nateglinide but not repaglinide inhibition. This finding is consistent with the idea that the binding site for nateglinide overlaps with that of tolbutamide and includes residue 1237. In contrast, the binding site for repaglinide does not include S1237 (it is unlikely that repaglinide could interact with FIG. 6. A: Structural formulas of compounds mentioned in the text. B: S1237 and not be affected by a mutation at this position, Superposition of repaglinide (red) and glibenclamide (green). C: Su- perposition of glibenclamide (green), tolbutamide (magenta), and because a serine to a tyrosine represents a rather large nateglinide (blue). change in side-chain volume).

DIABETES, VOL. 51, SEPTEMBER 2002 2793 INTERACTION OF REPAGLINIDE AND NATEGLINIDE WITH SUR1

Glibenclamide produced a reversible block of Kir6.2/ vations on the efficacy of repaglinide and nateglinide in SUR1[S1237Y]. In previous studies, it was not possible to vivo, or in the clinic, is uncertain. However, our data detect [3H]glibenclamide binding to this channel, a finding suggest that the structural determinants involved in nateg- that was suggested to be because the drug rapidly disso- linide binding overlap with those of sulphonylureas such ciates from SUR during the washing procedure of the as tolbutamide. In contrast, the properties of repaglinide binding assay (8). Our binding data provide additional binding and inhibition differ from those of the sulphonyl- support for this hypothesis. Thus, glibenclamide displaced ureas. This is of relevance in terms of other known [3H]repaglinide binding to the mutant channel with a much differences in the mechanism of action of repaglinide and lower potency than for the wild-type channel, consistent that of other insulin secretagogues. with a larger dissociation rate constant for glibenclamide In addition to closing KATP channels, sulphonylureas binding to Kir6.2/SUR1[S1237Y]. The lower binding affinity (e.g., tolbutamide) have been shown to interact directly of the mutant channel is also consistent with the fact that with the secretory machinery of ␤- and ␣-cells, thereby 2ϩ 1 ␮mol/l glibenclamide did not inhibit the channel com- stimulating Ca -dependent exocytosis of insulin and glu- pletely in patch-clamp experiments. Nateglinide, like tol- cacon, respectively (32,33). Likewise, tolbutamide stimulates direct exocytosis of somatostatin and growth butamide, inhibited wild-type KATP currents reversibly. In contrast, repaglinide inhibition of both wild-type and mu- hormone in vitro (34–36). This has led to the suggestion tant channels was not reversible. This suggests that the that a sulphonylurea binding protein exists on the secre- binding affinity of repaglinide is enhanced by interaction tory granules (37,38). It is of interest that nateglinide, but with additional residues in SUR1 and that this interaction not repaglinide, shares the ability to stimulate exocytosis is not disrupted by the S1237Y mutation. (34–36). It is possible, although unproven, that the shared The effect of the S1237Y mutation on nateglinide-in- characteristics of the binding sites for nateglinide and sulphonylureas, demonstrated in this study, help explain duced K channel inhibition could be due to either ATP these differences. reduced drug binding or an impaired ability of SUR1 to In conclusion, our data indicate that S1237 in SUR1 is a transduce drug binding into channel closure. However, prerequisite for the ability of tolbutamide and nateglinide because Kir6.2/SUR1[S1237Y] retains the ability to be to act as KATP channel inhibitors. In contrast, gliben- blocked fully by repaglinide and glibenclamide, the trans- clamide (partially) and repaglinide (wholly) interact with duction mechanism does not appear to be compromised regions of SUR1 that are distinct from S1237. by the mutation. (We assume that the transduction mechanism is common for all drugs used in this study.) Fur- thermore, the mutation impairs the ability of nateglinide to ACKNOWLEDGMENTS displace [3H]repaglinide binding, consistent with a re- We thank Tina Moeller Tagmose and Jesper Boeggild duced affinity for the former drug. Kristensen for synthesis and radiolabelling of repaglinide. The residue S1237 is positioned in the intracellular loop F.M.A. is the GlaxoSmithKline Royal Society Research between TM segments 15 and 16. The importance of this Professor. residue for binding and channel inhibition by both tolbutamide and nateglinide strongly argues for an intracellular binding site for these drugs. Nateglinide has been pro- REFERENCES 1. Ashcroft FM, Rorsman P: Electrophysiology of the pancreatic beta-cell. posed to act from the extracellular side and not to be Prog Biophys Mol Biol 54:87–143 1987 internalized into the ␤-cells (30,31). The results presented 2. Ashcroft FM, Gribble FM: ATP-sensitive Kϩ channels and insulin secre- here, however, suggest that the binding site for nateglinide tion: their role in health and disease. Diabetologia 42:903–919, 1999 resides on the intracellular side of SUR1 and shares at 3. Trube G, Rorsman P, Ohno-Shosaku T: Opposite effects of tolbutamide and diazoxide on the ATP-dependent Kϩ channel in mouse pancreatic beta- least one point of interaction with the binding site for cells. Pflugers Arch 407:493–499, 1986 tolbutamide. 4. Ashcroft FM: Adenosine 5Ј-triphosphate-sensitive potassium channels. Structural comparison. The modeling studies demon- Annu Rev Neurosci 11:97–118, 1988 strate that glibenclamide and repaglinide can be fittoa 5. Inagaki N, Gonoi T, Clement JP, Namba N, Inazawa J, Gonzalez G, common pharmacophore model, whereas nateglinide, tol- Aguilar-Bryan L, Seino S, Bryan J: Reconstitution of IKATP: an inward rectifier subunit plus the sulfonylurea receptor. Science 270:1166–1170, butamide, and glibenclamide can be fit to another model 1995 that is distinct from that of glibenclamide/repaglinide. The 6. Clement JP, Kunjilwar K, Gonzalez G, Schwanstecher M, Panten U, conformations of glibenclamide in the two models differ Aguilar-Bryan L, Bryan J: Association and stoichiometry of K(ATP) chan- slightly, but we cannot rule out the possibility that the nel subunits. Neuron 18:827–838, 1997 7. Aguilar-Bryan L, Nichols CG, Wechsler SW, Clement JP 4th, Boyd AE 3rd, large number of conformational degrees of freedom en- Gonzalez G, Herrera-Sosa H, Nguy K, Bryan J, Nelson DA: Cloning of the able glibenclamide to adopt a conformation that simulta- beta cell high-affinity sulfonylurea receptor: a regulator of insulin secre- neously fits both models. Most importantly, the structural tion. Science 268:423–426, 1995 investigations illustrate that the acidic moiety is the only 8. Ashfield R, Gribble FM, Ashcroft SJ, Ashcroft FM: Identification of the pharmacophore group shared by all four drug molecules. high-affinity tolbutamide site on the SUR1 subunit of the K(ATP) channel. Diabetes 48:1341–1347, 1999 They further reveal that the amide part of glibenclamide is 9. Gribble FM, Tucker SJ, Seino S, Ashcroft FM: Tissue specificity of shared with repaglinide, whereas the sulphonylurea part sulfonylureas: studies on cloned cardiac and ␤-cell K(ATP) channels. of glibenclamide is shared with both tolbutamide and Diabetes 47:1412–1418, 1998 nateglinide. This provides additional support for the idea 10. Inagaki N, Gonoi T, Clement JP, Wang CZ, Aguilar-Bryan L, Bryan J, Seino S: A family of sulfonylurea receptors determines the pharmacological that nateglinide shares (at least in part) a common site of properties of ATP-sensitive Kϩ channels. Neuron 16:1011–1017, 1996 interaction on SUR1 with sulphonylureas. 11. Isomoto S, Kondo C, Yamada M, Matsumoto S, Higashiguchi O, Horio Y, Pharmacological relevance. The impact of these obser- Matsuzawa Y, Kurachi Y: A novel sulfonylurea receptor forms with BIR

2794 DIABETES, VOL. 51, SEPTEMBER 2002 A.M.K. HANSEN AND ASSOCIATES

(Kir6.2) a smooth muscle type ATP-sensitive Kϩ channel. J Biol Chem nateglinide: study on cardiovascular and beta-cell K(ATP) channels. 271:24321–24324, 1996 J Pharmacol Exp Ther 291:1372–1379, 1999 12. Garrino MG, Schmeer W, Nenquin M, Meissner HP, Henquin JC: Mecha- 26. Dorschner H, Brekardin E, Uhde I, Schwanstecher C, Schwanstecher M: nism of the stimulation of insulin release in vitro by HB 699, a benzoic acid Stoichiometry of sulfonylurea-induced ATP-sensitive potassium channel derivative similar to the non-sulphonylurea moiety of glibenclamide. closure. Mol Pharmacol 55:1060–1066, 1999 Diabetologia 28:697–703, 1985 27. Fuhlendorff J, Rorsman P, Kofod H, Brand CL, Rolin B, MacKay P, Shymko 13. Loffler-Walz C, Hambrock A, Quast U: Interaction of K(ATP) channel R, Carr RD: Stimulation of insulin release by repaglinide and glibenclamide modulators with sulfonylurea receptor SUR2B: implication for tetramer involves both common and distinct processes. Diabetes 47:345–351, 1998 formation and allosteric coupling of subunits. Mol Pharmacol 61:407–414, 28. Fujita T, Seto Y, Kondo N, Kato R: Studies on the N-(trans-4-isopropylcy- 2002 clohexyl)-carbonyl-D-phenylalanin e (A-4166)receptor in HIT T-15 cells: 14. Mikhailov MV, Ashcroft SJH: Interactions of the sulfonylurea receptor 1 displacement of [3H]glibenclamide. Biochem Pharmacol 52:407–411, 1996 subunit in the molecular assembly of beta-cell K-ATP channels. J Biol 29. Ikenoue T, Akiyoshi M, Fujitani S, Okazaki K, Kondo N, Maki T: Hypogly- Chem 275:3360–3364, 2000 caemic and insulinotropic effects of a novel oral antidiabetic agent, 15. Mikhailov MV, Mikhailova EA, Ashcroft SJH: Molecular structure of the (-)-N-(trans-4-isopropylcyclohexanecarbonyl)-D-phenylalanine (A-4166). glibenclamide binding site of the beta-cell K-ATP channel. FEBS Lett Br J Pharmacol 120:137–145, 1997 30. Malaisse-Lagae F, Malaisse WJ: Fate of 3H- and 14C-labelled A-4166 in 499:154–160, 2001 pancreatic islets. Acta Diabetol 33:298–300, 1996 16. Reimann F, Proks P, Ashcroft FM: Effects of mitiglinide (S 21403) on 31. Hu S, Wang S, Dunning BE: Glucose-dependent and glucose-sensitizing Kir6.2/SUR1, Kir6.2/SUR2A and Kir6.2/SUR2B types of ATP-sensitive po- insulinotropic effect of nateglinide: comparison to sulfonylureas and tassium channel. Br J Pharmacol 132:1542–1548, 2001 repaglinide. Int Jnl Experimental Diabetes Res 2:63–72, 2001 17. Grell W, Hurnaus R, Griss G, Sauter R, Rupprecht E, Mark M, Luger P, Nar 32. Gromada J, Dissing S, Kofod H, Frokjaer-Jensen J: Effects of the hypogly- H, Wittneben H, Muller P: Repaglinide and related hypoglycemic benzoic caemic drugs repaglinide and glibenclamide on ATP-sensitive potassium- J Med Chem acid derivatives. 41:5219–5246, 1998 channels and cytosolic calcium levels in beta TC3 cells and rat pancreatic 18. Hu SL, Wang SY, Fanelli B, Bell PA, Dunning BE, Geisse S, Schmitz R, beta cells. Diabetologia 38:1025–1032, 1995 Boettcher BR: Pancreatic beta-cell K-ATP channel activity and membrane- 33. Bokvist K, Hoy M, Poulsen CR, Buschard K, Gromada J: A4166, but not binding studies with nateglinide: a comparison with sulfonylureas and repaglinde stimulate Ca2ϩ-evoked KATP-channel independent, secretion repaglinide. J Pharmacol Exp Ther 293:444–452, 2000 in rat pancreatic alpha- and beta-cells (Abstract). Diabetologia 41:A139, 19. Dabrowski M, Wahl P, Holmes WE, Ashcroft FM: Effect of repaglinide on 1998 cloned beta cell, cardiac and smooth muscle types of ATP-sensitive 34. Bokvist K, Hoy M, Buschard K, Holst JJ, Thomsen MK, Gromada J: potassium channels. Diabetologia 44:747–756, 2001 Selectivity of prandial glucose regulators: nateglinide, but not repaglinide, 20. Mohamadi F, Richards NGJ, Guida WC, Liskamp R, Lipton M, Caufield C, accelerates exocytosis in rat pancreatic A-cells. Eur J Pharm 386:105–111, Chang G, Hendrickson T, Still WC: MacroModel: an integrated software 1999 system for modeling organic and bioorganic molecules using molecular 35. Hoy M, Olsen HL, Bokvist K, Buschard K, Barg S, Rorsman P, Gromada J: mechanics. J Comput Chem 11:440–467, 1990 Tolbutamide stimulates exocytosis of glucagon by inhibition of a mito- 21. Halgren TA, Nachbar RB: Merck molecular-force field. IV. Conformational chondrial-like ATP-sensitive Kϩ (KATP) conductance in rat pancreatic energies and geometries for mmff94. J Comput Chem 17:587–615, 1996 A-cells. J Physiol 527:109–120, 2000 22. Still WC, Tempczyk A, Hawley RC, Hendrickson T: Semianalytical treat- 36. Fujitani S, Ikenoue T, Akiyoshi M, Maki T, Yada T: Somatostatin and ment of solvation for molecular mechanics and dynamics. J Am Chem Soc insulin secretion due to common mechanisms by a new hypoglycemic 112:6127–6129, 1990 agent, A-4166, in perfused rat pancreas. Metabolism 45:184–189, 1996 23. Jones G, Willett P, Glen RC: A genetic algorithm for flexible molecular 37. Eliasson L, Renstrom E, Ammala C, Berggren PO, Bertorello AM, Bokvist overlay and pharmacophore elucidation. J Comput Aided Mol Des 9:532– K, Chibalin A, Deeney JT, Flatt PR, Gabel J, Gromada J, Larsson O, 549, 1995 Lindstrom P, Rhodes CJ, Rorsman P: PKC-dependent stimulation of 24. Sunaga Y, Gonoi T, Shibasaki T, Ichikawa K, Kusama H, Yano H, Seino S: exocytosis by sulfonylureas in pancreatic beta cells. Science 271:813–815, The effects of mitiglinide (KAD-1229), a new anti-diabetic drug, on 1996 ATP-sensitive Kϩ channels and insulin secretion: comparison with the 38. Ozanne SE, Guest PC, Hutton JC, Hales CN: Intracellular localization and sulfonylureas and nateglinide. Eur J Pharm 431:119–125, 2001 molecular heterogeneity of the sulphonylurea receptor in insulin-secreting 25. Hu S, Wang S, Dunning BE: Tissue selectivity of antidiabetic agent cells. Diabetologia 38:277–282, 1995

DIABETES, VOL. 51, SEPTEMBER 2002 2795