Expression of Voltage-Gated Potassium Channels in Human and Rhesus Pancreatic Islets Lizhen Yan,1 David J

Expression of Voltage-Gated Potassium Channels in Human and Rhesus Pancreatic Islets Lizhen Yan,1 David J. Figueroa,2 Christopher P. Austin,3,4 Yuan Liu,5 Randal M. Bugianesi,1 Robert S. Slaughter,1 Gregory J. Kaczorowski,1 and Martin G. Kohler1

Voltage-gated potassium channels (Kv channels) are urea receptor SUR1, sets the ␤-cell resting membrane involved in repolarization of excitable cells. In pancre- potentials (Em) under low plasma glucose conditions (3,4). atic ␤-cells, prolongation of the action potential by Elevated plasma glucose concentration results in an in- block of delayed rectifier potassium channels would be crease in metabolic activity, which leads to closure of KATP expected to increase intracellular free calcium and to channels and to membrane depolarization (5,6). Voltage- promote insulin release in a glucose-dependent manner. gated calcium channels (Ca channels) then become acti- However, the specific Kv channel subtypes responsible vated, and the resultant rise in intracellular Ca2ϩ triggers for repolarization in ␤-cells, most importantly in hu- mans, are not completely resolved. In this study, we insulin secretion. Sulfonylureas, widely used insulin secre- have investigated the expression of 26 subtypes from Kv tagogues, bind to the SUR1 receptor and block the KATP subfamilies in human islet mRNA. The results of the channel, causing insulin secretion in the absence of glu- RT-PCR analysis were extended by in situ hybridization cose metabolism (7,8). ␤ and/or immunohistochemical analysis on sections from KCa in -cells consists of at least two different compo- human or Rhesus pancreas. Cell-specific markers were ␤ nents. There is a large conductance KCa channel in -cells, used to show that Kv2.1, Kv3.2, Kv6.2, and Kv9.3 are with no obvious physiological function (9). In nondissoci- ␤ expressed in -cells, that Kv3.1 and Kv6.1 are expressed ated ␤-cells, a second type of K current has been in ␣-cells , and that Kv2.2 is expressed in ␦-cells. This Ca study suggests that more than one Kv channel subtype described (10). This current is linked to depolarization- might contribute to the ␤-cell delayed rectifier current induced rhythmic electrical activity of ␤-cells, important and that this current could be formed by heterotetra- for insulin secretion (11). mers of active and silent subunits. Diabetes 53: The remaining potassium current is generated by Kv 597–607, 2004 channels that produce either a fast transient current, IA, or a slow inactivating, delayed rectifying current, IDR (12–14). ␤ Both currents exist in -cells, with IDR being the major contributor to the repolarization of these cells. Thus, he role of potassium channels in excitation- blockage of IDR should enhance Ca influx and therefore secretion coupling is well established (1). In lead to an increase in insulin secretion as has been ␤ pancreatic -cells, insulin secretion is modulated previously reported (15–17). Because IDR does not open Tby the activity of different ionic currents. Among until the membrane is depolarized above a threshold level these are the three main potassium currents found in of ca Ϫ20 mV, its activation would be glucose dependent. ␤ -cells: the ATP-sensitive (KATP), calcium-activated (KCa), Therefore, pharmacological interference with this mecha- and voltage-gated (Kv) currents. Each has a functional role nism may provide a novel way to treat type 2 diabetes at different stages in the process of glucose-induced insu- without causing the hypoglycemic adverse effect of sulfo- lin secretion (2). nylureas (18,19). IDR has also been shown to be part of the KATP, consisting of inward rectifier Kir6.2 and sulfonyl- signaling pathway of glucagon-like peptide-1–induced glucose-dependent insulin secretion (20). The function of IDR ␦ ␤ From the 1Department of Ion Channels, Merck Research Laboratories, Rah- currents in -cells may be similar to that of -cells because way, New Jersey; the 2Department of Molecular and Investigative Toxicology, action potential initiation is dependent on depolarization Merck Research Laboratories, West Point, Pennsylvania; the 3Department of through metabolism-dependent blockage of K .In Neuroscience, Merck Research Laboratories, West Point, Pennsylvania; ATP the 4National Human Genome Research Institute, Bethesda, Maryland; and the ␣-cells, the opening of Na channels apparently initiates the 5Department of Bioinformatics, Merck Research Laboratories, West Point, action potential, but the IDR may still be involved in Pennsylvania. Address correspondence and reprint requests to Dr. Lizhen Yan or Dr. repolarization (21). Martin G. Kohler, RY80N-C31, Department of Ion Channels, Merck Research Kv channels belong to the six-transmembrane (TM) Laboratories, Rahway, NJ 07065. E-mail: [email protected] or martin_ family of K channels, where Kv1 to Kv11 subfamilies exist, [email protected]. Received for publication 13 April 2003 and accepted in revised form 14 although Kv7 is only found in Aplysia (22,23). Members of November 2003. the Kv1 to Kv4 subfamilies form tetrameric functional L.Y. and D.J.F. contributed equally to this article. channels, homomultimers or heteromultimers, usually DAPI, 4,6-diamidino-2-phenylindole; Em, membrane potential; IA, fast transient current; IDR, delayed rectifying current; IHC, immunohistochemistry; with members from the same subfamily. Members of the ISH, in situ hybridization; KATP current, ATP-sensitive potassium current; KCa Kv5 to Kv11 families code for “silent subunits” that do not current, calcium-activated potassium current; Kv current, voltage-gated potassium current; TM, transmembrane. express as functional homomultimers. In heterologous © 2004 by the American Diabetes Association. expression systems, silent subunits can coassemble with

DIABETES, VOL. 53, MARCH 2004 597 Kv CHANNELS IN PRIMATE ISLETS

TABLE 1 PCR primers used for RT-PCR ampliﬁcation of Kv channels Subtype Accession no. Sense primer (5Ј to 3Ј) Antisense primer (5Ј to 3Ј) Kv1.1 L02750 CATCTGGTTCTCCTTCGAGC GTTAGGGGAACTGACGTGGA Kv1.2 L02752 TCCGGGATGAGAATGAAGAC TTGGACAGCTTGTCACTTGC Kv1.3 M55515 GTTCTCCTTCGAACTGCTGG CTGAAGAGGAGAGGTGCTGG Kv1.4 M55514 CCCCAGCTTTGATGCCATCTTG TGAGGATGGCAAAGGACATGGC Kv1.5 M55513 TGCGTCATCTGGTTCACCTTCG TGTTCAGCAAGCCTCCCATTCC Kv1.6 X17622 TCAACAGGATGGAAACCAGCCC CTGCCATCTGCAACACGATTCC Kv1.7 AJ310479 TGCCCTTCAATGACCCGTTCTTC AAGACACGCACCAATCGGATGAC Kv2.1 L02840 TACAGCCTCGACGACAACG ACCACGCGGCGGACATTCTG Kv2.2 U69962 AACGAAGAACTGAGGCGAGAG ACTCCGCCTAAGGGTGAAAC Kv3.1 S56770 AACCCCATCGTGAACAAGACGG TCATGGTGACCACGGCCCA Kv3.2 AI363404 CTGCTGCTGGATGACCTACC TGTGCCATTGATGACTGGTT Kv3.3 AF055989 TTCTGCCTGGAAACCCATGAGG TGTTGACAATGACGGGCACAGG Kv3.4 M64676 TTCAAGCTCACACGCCACTTCG TGCCAAATCCCAAGGTCTGAGG Kv4.1 AJ005898 ATCTCGAGGAGATGAGGTTC TTCTTTCGGTCCCGATAC Kv4.3 AF048712 TGGCTTCTTCATCGCTGTCTCG CCGAAGATCTTCCCTGCAATCG Kv4.4 NM_012283 AGCCAAGAAGAACAAGCTG AGGAAGTTTAGGACATGCC Kv5.1 AF033382 TCCACATGAAGAAGGGCATCTGC TCACGTAGAAGGGGAGGATG Kv6.1 AF033383 TGCACCAACTTCGACGACATCC GGAACTCCAGGGAGAACCAGCC Kv6.2 AJ0111021 AAGCTCTTCGCCTGCGTGTC CAGCAGCAGCGACACGTAGAAC Kv6.3 NM_172347 ATGCCCATGCCTTCCAGAGA AGAGCTGCACGATCTCCTCG Kv8.1 AF167082 TTCCACAGCTGCCCGTATCTTTG TTTTGCCTGTGGTGGTGTCTGG Kv9.1 AF043473 TTTGAGGACTTGCTGAGCAGCG TTGCTCCAGGCACACCAACAAG Kv9.2 XM_043106 GTACTGGGGCATCAACGAGT CCACGGAGAGGTAGAGCAAG Kv9.3 AF043472 CTCTGTGGGCATTTCCATTT AGAAACAGGCACAAACACCC Kv10.1 AF348982 GCTTGCCCGTCACTTCATTGGTC TTCTTCCAGGCACTGTGATAGGA Kv11.1 AF348983 AGCCATGCTCAAACAGAGTG CTCCTCGTAGTCGTCGCACA

Kv2 and Kv3 subunits and modulate the biophysical char- subtype-specific primer pair for each of the 26 members of the Kv1–Kv11 acteristics of the latter subunits (24–27). There are several families. To obtain efficient and specific primer pairs, we used the Vector NTI difficulties that obscure the correlation of any particular program (Informax, Frederick, MD) to select sequences. Each primer se- Kv subunit with a specific physiological function: the high quence then was submitted to a basic local alignment search tool search degree of sequence homology results in many Kv channels against GenBank to ensure specificity of the selected sequence. Specific having similar pharmacological and biophysical proper- primer pairs were then used to amplify Kv channel subtypes from human fetal ties, and most excitable cells express more than one Kv brain cDNA. Finally, the most effective primer pairs for each subtype were channel gene. used to study the expression of Kv channels in human islets (Table 1). Antibodies. A rabbit polyclonal antibody for Kv1.6 was raised against the Previous studies have identified Kv2.1 and Kv3.2 in peptide RRSSYLPTPHRAYAEKRM, corresponding to residue 509–526 of the rodent ␤-cells and insulinoma cells (16,28–30), and block rat Kv1.6 (34). Human Kv1.6 shares 17 of 18 amino acid residues with the rat of Kv2.1 has been implicated in eliciting glucose-depen- channel in this region. Rabbit polyclonal antibodies against Kv2.1 and Kv3.2 dent insulin secretion (16,31,32). While Kv2.1 has been proteins were purchased from Alomone Labs (Jerusalem, Israel). The Kv2.1 detected in human islets (33), no studies have yet been antibody was raised against the peptide HMLPGGGAHGSTRDQSI, corre- ␤ sponding to residue 837–853 of rat Kv2.1. Human Kv2.1 shares 15 of 17 amino attempted in human or primate -cells to define the acids with this region of the rat channel. The Kv3.2 antibody was raised molecular components of IDR. Although a molecular basis against the peptide DLGGKRLGIEDAAGLGGPDGK(C), corresponding to res- for the A-type current has been reported in ␣- and ␦-cells idues 184–204 of rat Kv3.2. The human and rat Kv3.2 have 19 identical amino acids in this peptide sequence. Antibodies for Kv2.1 and Kv3.2 detected only (21), there is no information on the IDR in these cells. In this study, we used RT-PCR to analyze the expression a single band in Western blots of the cognate channels expressed in HEK 293 cells. of 26 Kv channel genes in human islets. Cell-type specific RT-PCR. Human pancreatic islets were obtained from the University of expression of 11 Kv subtypes was further determined by in Alberta (Edmonton, AB, Canada). The islets were purified based on staining situ hybridization or immunohistochemistry. All data, with the ␤-cell–specific dye diphenylthiocarbazone (Sigma), as previously taken together, suggest that closely related Kv channel described (35). The total RNA was isolated using TRI Reagent (Molecular subtypes are distributed among different cell types in Research Center, Cincinnati, OH) according to the manufacturer’s instruction. Because most of the human Kv channel genes are intronless, we treated the primate islets. In addition, we provide evidence that total RNA with DNaseI (Ambion, Austin, TX) to eliminate traces of genomic pancreatic ␤-cells express both silent and functionally DNA. A control PCR reaction was performed with ␤-actin primers (forward: active Kv channel subtypes. Some heteromeric combina- 5Ј-GCCCTTTCTCACTGGTTCTC-3Ј; reverse: 5Ј-CTTTACACCAGCCTCAT tion of these subtypes might be the underlying molecular GGC-3Ј) located on an intron region to verify the absence of genomic DNA. correlate of I . The DNaseI-digested RNA was transcribed into cDNA using SensiScript DR Reverse Transcriptase from Qiagen (Valencia, CA), per the manufacturer’s instruction. Human fetal brain poly Aϩ RNA was purchased from BD RESEARCH DESIGN AND METHODS Biosciences Clontech (Palo Alto, CA). The RNA was transcribed into cDNA Design and optimization of RT-PCR primers for Kv1 to Kv11 family with the Omniscript Reverse Transcriptase Kit from Qiagen following the subtypes. To establish the profile of Kv channel subtypes in islets, we manufacturer’s instructions. The cDNA was used as templates for the ampli- performed RT-PCR amplification on RNA extracted from human islets using a fication of individual channel subtypes. Approximately 100 ng of total RNA

598 DIABETES, VOL. 53, MARCH 2004 L. YAN AND ASSOCIATES

TABLE 2 DNA probes used for ISH Subtype Sequence (5Ј to 3Ј) hKv2.1 CTCAAAGTTGAACGCTATTGCTGTGTGTTTCTCAGGAGACC TGACCAATCATTCCCTGTAGCTGTCTAACAGTGGAATCCATCC hKv2.2 GGACTGTCGGTGGCATTGTCAGACTGCAAAGGACTATGTAGC CCATCCCTGGCAGCAGGTCCTTTCTCTCTGGATAGAGTTAG hKv3.1 TCGGCGTCCGCGTCGTCGGCGCTGTTGTCCAGAGGAG CTCCCGGTAGTAGCGCACTTGCGTGCCATTGCGAACG hKv3.2 GCCAGGCGTGTTCCAGGCAGGGTCTTGAGGGTG TTCGAAGACCCCTACTCGTCCAGAGCCGCCAGGTTTA hKv4.1 CCTGGGTCACCGTCTTGTTGCTAATATGGATGAAGCC GTCGGTGAGGAGGAAGCAGGCTCGGTCCCGGCTATA hKv6.1 GTGAAGGCATCTCGAGGAGATGAGGTTC GAAGGAATTCTTCTACGATGCTGACTCA hKv6.2 TGAGGATGTCGTCGAAGTTGGTGCAGGCCTTGAGCTG GATGAGGAACTGCGCCCTCCGGAACATCACCCTCT hKv9.2 ACTGGGCCACAAGCACTAGAATAGCGTACACGATGGAC GGCTGCCCATCGAACTTGGAGGCGTCGTTGTAGAAGGC hKv9.3 GAGGAAACCGCAGGAGGGTGCTTTGGTCAACAGACTGC GGAGCTGACCAAATCGCAGTGTGTCAAACTTCTCCAGC was used in each reaction. A blank reaction was used as a control. The CA) and sequenced to verify their identity. To prepare riboprobes for ISH, the reaction conditions were as follows: start with 95°C for 15 min to activate the Kv3.3 and Kv9.2 containing vectors were linearized with the restriction HotStarTaq DNA Polymerase (Qiagen), then denature at 94°C for 30 s, enzymes SpeI and NotI, respectively, to create template DNA. Biotinylated annealing at 56°C for 30 s, followed by an extension at 72°C for 1 min. The sense and antisense riboprobes were then generated by in vitro transcription total number of cycles was 35. The amplification was followed by a 10-min using Biotin RNA Labeling Mix (Roche Molecular Biochemicals). extension at 72°C. The primers for each Kv channel subtype are listed in Table Cryostat sections (8 ␮m) of human pancreas (the National Disease 1. The sequences for the insulin primers are the following: sense primer Research Interchange) and rhesus pancreas tissue, obtained under the ap- 5Ј-CCAGCCGCAGCCTTTGTGA-3Ј, antisense primer 5Ј-GCTGGTAGAGGGAG proval of the Merck Research Laboratories Institutional Animal Care and Use CAGAT-3Ј. The sequences for the trypsin II primers are the following: sense Committee, were thaw mounted on SuperFrost plus slides (Fisher Scientific) primer 5Ј-GCCCCCTTTGATGATGATG-3Ј, antisense primer 5Ј-ACACGCGG and fixed with 4% paraformaldehyde. To improve the signal strength, a Ј GAATTGATGAC-3 . The sequences for the Kir6.2 primers are the following: cocktail mixture of two labeled oligonucleotide probes (listed in Table 2) sense primer 5Ј-AAGAAGTGAAGTGGGACC-3Ј, antisense primer 5Ј-GTTGC specific for a particular Kv channel subunit was used at a final concentration CTTTCTTGGACAC-3Ј. of 2 pmol/ml each. The ISH conditions were the same as those previously In situ hybridization, immunohistochemistry, and double-label com- described (36). Bound probes were detected using the TSA direct red FISH bined in situ hybridization/immunohistochemistry. For in situ hybridiza- tyramide amplification kit (PerkinElmer Life Sciences, Boston, MA) according tion (ISH), oligonucleotide probes specific for human Kv2.1, Kv2.2, Kv3.1, to the manufacturer’s instructions. ISH using mRNA probes, immunohisto- Kv6.1, Kv6.2, and Kv9.3 (Table 2) were end labeled with biotin-16-dUTP chemistry (IHC) using antibodies, and combined double-label ISH/IHC were (Roche Molecular Biochemicals, Indianapolis, IN). The digoxigenin oligonu- performed as previously described (37). ISH and IHC expression experiments cleotide tailing kit (Roche Molecular Biochemicals) was used according to the were carried out on human and Rhesus pancreatic sections. The expression of manufacturer’s protocol, except for replacement of digoxigenin-dUTP with Kv2.1, 2.2, 3.1, 3.2, 3.3, 6.1, 6.2, 9.2, and 9.3 subunits were examined in sections biotin-16-dUTP. For Kv3.3 and Kv9.2, riboprobes instead of oligo probes were from both species, and in all cases, expression patterns in Rhesus and human used for ISH. A 298-bp fragment from the 3Ј-untranslated region of the human were consistent. Double-staining cell identification experiments were carried Kv3.3 cDNA insert was prepared by RT-PCR amplification from human fetal out on the specimens that gave the best signal. Kv1.6 and Kv4.1 were only brain cDNA, using a Kv3.3-specific primer pair (Table 1). Similarly, a 543-bp monitored in Rhesus. fragment from the 3Ј-untranslated region of the human Kv9.2 cDNA was Cell identification markers were antibodies specific for glucagon (Dako, amplified, using a Kv9.2-specific primer pair (Table 1). The reaction conditions Carpinteria, CA), insulin (Zymed Laboratories, South San Francisco, CA), and were the same as described above for subtype RT-PCR. The PCR fragments somatostatin (Dako). Matched preimmune sera or nonimmune control sera were subcloned into the plasmid vector pCRII-TOPO (Invitrogen, Carlsbad, were used as negative controls for IHC. All antibodies were used at the

FIG. 1. RT-PCR amplification of Kv channel subtypes from RNA of human fetal brain and pancreatic islets. A: Human fetal brain poly A؉ RNA was subjected to RT-PCR amplification using primers shown in Table 1 as described in RESEARCH DESIGN AND METHODS. B: RNA extrac- tion from pancreatic islets was performed and total RNA was reverse transcribed into cDNA. PCR amplifications of Kv channel subtypes in islets were performed under the same conditions as described in A. Results from the elec- trophoresis of the PCR products that were separated on 1% agarose gels and stained with ethidium bromide are presented. On the left side of each image are the DNA markers (800 ng/lane; Bio-Rad Laboratories, Hercules, CA) of different sizes (bp). The top of each image labels the subtype of Kv channel represented by each lane. All PCR products showed the expected molecular sizes; some were cloned in bacterial vectors and sequenced to verify identity (data not shown). PCR reactions with water instead of cDNA were performed as a control (data not shown.)

DIABETES, VOL. 53, MARCH 2004 599 Kv CHANNELS IN PRIMATE ISLETS

FIG. 2. Kv2.1 channel is expressed in pancreatic ␤-cells and Kv2.2 channel is expressed in ␦-cells. In all sections, cell nuclei were stained with DAPI and are shown in blue. A: ISH for Kv2.1 mRNA-positive cells (red) can be seen within an islet from Rhesus pancreas. B: IHC with insulin detects ␤-cells (green) within the same islet as seen in panel A. C: Colocalization (yellow) of Kv2.1 mRNA and insulin within the same section. D: Results from a double-staining ISH/IHC experiment including Kv2.1 mRNA (red) and glucagon protein (green) in a human islet. The lack of yellow indicates no coexpression of Kv2.1 mRNA with glucagon. E: IHC of Kv2.1 protein visualized with Texas red (red) labels cells within a Rhesus islet. F: Kv2.2 mRNA-containing cells (red) are present within a Rhesus islet. G: Insulin label of ␤-cells (green) in the same region of the islet. H: Lack of colocalization (yellow) of Kv2.2 mRNA and insulin protein. I: Kv2.2 mRNA-positive cells (red) are present within a region of a Rhesus islet. J: Somatostatin-containing pancreatic ␦-cells (green) within the same section of the islet as in panel I. K: Kv2.2 mRNA colocalizes (yellow) with somatostatin.

600 DIABETES, VOL. 53, MARCH 2004 L. YAN AND ASSOCIATES

FIG. 3. Expression of Kv3.1 and Kv3.2 channels in pancreatic ␣- and ␤-cells, respectively. In all sections, cell nuclei were stained with DAPI and are illustrated in blue. A: Kv3.1 mRNA-containing cells (red) in a section of a human islet. B: Insulin protein-containing cells (green) in the same section used in panel A. C: Lack of colocalization (no yellow color) of Kv3.1 mRNA and insulin protein. D: Kv3.1 mRNA-positive cells (red) within a Rhesus islet. E: Glucagon-positive ␣-cells (green) are present within the same islet. F: Kv3.1 mRNA colocalizes with glucagon (yellow). G: Kv3.2 mRNA-containing cells (red) highlight an islet in a section from a Rhesus pancreas. H: Insulin immunoreactive cells (green) shown in the same islet. I: Colocalization of Kv3.2 mRNA and insulin protein (yellow). J: IHC of Kv3.2 protein visualized with fluorescein isothiocyanate (green) labels cells within a Rhesus islet. K: Kv2.1 mRNA-positive cells (red) shown within the same islet. L: Kv3.2 protein colocalizes with Kv2.1 mRNA in the same cells (yellow). manufacturer’s specified dilutions and incubated on the sections for2hat with 4,6-diamidino-2-phenylindole (DAPI; Molecular Probes, Eugene, OR). room temperature following the ISH procedures. Bound antibodies were Digital acquisition and image reassembly were carried out using a MicroMax detected using fluorescein isothiocyanate–conjugated donkey (MultipleLabel) CCD camera (Princeton Instruments, Princeton, NJ) and Metamorph imaging IgG (Jackson Immunoresearch, West Grove, PA). Nuclei were counterstained software (Universal Imaging, West Chester, PA).

DIABETES, VOL. 53, MARCH 2004 601 Kv CHANNELS IN PRIMATE ISLETS

FIG. 4. Kv6.1 channel is expressed in islet ␣-cells, and Kv6.2 channel is expressed in islet ␤-cells. In all sections, cell nuclei were stained with DAPI and are illustrated in blue. A: Kv6.1 mRNA-containing cells (red) are found within a human islet. B: Insulin-positive cells (green) in the same section as in panel A highlight different islet cells. C and D: Microscopic analysis reveals lack of colocalization (no yellow color) between Kv6.1 mRNA and insulin. E: Kv6.1 mRNA-containing cells (red) in a section of a human islet. F: Glucagon protein-containing cells (green) in the same section used in E. G and H: Colocalization (yellow) of Kv6.1 mRNA and glucagon protein was detected. I: Kv6.2 mRNA-containing cells (red) in a section of a Rhesus islet. J: Insulin protein-containing cells (green) in the same section used in I. K and L: Colocalization (yellow) of Kv6.2 mRNA and insulin protein.

RESULTS Kv3.4, Kv4.4, and Kv9.1 were only found in one of three There are 17 channel subtypes detected in human human islet preparations. Kv4.4 and Kv9.1 appear as faint islets by RT-PCR. The level of effectiveness of the Kv bands compared with brain and may be due to nonislet channel subtype-specific primers was tested by PCR using tissue in the one preparation. However, Kv3.4 was also human fetal brain cDNA as templates. In all cases, a PCR found in only one of three human islet cDNA preparations, product was visible on a 1% agarose gel (Fig. 1A). The but as a more prominent band. This channel is usually expression profile for Kv channels was determined in at responsible for an A-type potassium current and has been least two different human islet cDNA preparations. PCR reported to be in ␤- and ␦-cells from rat (21). This fragments for Kv1.3, 1.6, 1.7, 2.1, 2.2, 3.1, 3.2, 3.3, 4.1, 4.4, expression profile could be a natural variation within the 6.1, 6.2, 9.1 9.2, 9.3, 10.1, and 11.1 are identified as being human population or a marker for a possible disease state. present in the pancreatic islet preparation (Fig. 1B and The expression of Kv1.7 was found in all three human Table 3). Probes for Kir6.2, representative of islet RNA, preparations, but the band was weaker than the signal in and trypsin, representative of acinar tissue RNA, indicate brain and the signal of the other Kv channels in islets. that there is some nonislet RNA present in the human islet These results were confirmed in a separate experiment preparation used for Fig. 1. where Kv1.7 and Kv1.6 were tested in a tissue panel

602 DIABETES, VOL. 53, MARCH 2004 L. YAN AND ASSOCIATES

FIG. 5. Expression of Kv9.3, but not Kv9.2, channels in pancreatic ␤-cells. In all sections, cell nuclei were stained with DAPI and are illustrated in blue. A: Kv9.2 mRNA-containing cells (red) in a section of a human islet. B: Insulin immunoreactive cells (green) in the same section used in panel A. C: Lack of colocalization (no yellow color) of Kv9.2 mRNA and insulin protein. D: Kv9.3 mRNA-containing cells (red) within a Rhesus islet. E: Insulin-positive cells (green) in the same section used in D. F: Significant colocalization (yellow) of Kv9.3 mRNA and insulin protein was detected. G: Kv9.3 mRNA-containing cells (red) within an islet. H: Glucagon protein-containing cells (green) in the same section used in G. I: Lack of colocalization (no yellow color) of Kv9.3 mRNA and glucagon protein was detected. including islets, brain, and skeletal muscle among others, Because Kv2.1 and Kv3.2 have been reported to exist in where Kv1.7 was most prominent in skeletal muscle, ␤-cells and in insulin-secreting cell lines (16,28), the cellu- confirming a previous report (38). While RT-PCR is not lar presence of these two channel subtypes was deter- quantitative, these experiments suggest that the expres- mined first. Pancreatic sections were probed for Kv2.1 sion of Kv1.7 is lower in islets than for other Kv channels mRNA with labeled antisense oligonucleotides by ISH observed in this tissue. (Table 2) and with an insulin-specific antibody by IHC in a Kv2.1 and Kv3.2 colocalize with insulin in ␤-cells. double-staining experiment. Kv2.1 mRNA-positive cells Pancreatic islets are composed primarily of three cell located within an islet (Fig. 2A) exclusively colocalize with types (i.e., ␣-, ␤-, and ␦-cells); however, vascular endothe- the insulin-containing cells, which confirms the expression lial cells (39) and interneurons (40) are found in islets as of Kv2.1 in ␤-cells (Fig. 2B and C). Consistent with this well. To identify the presence of channel subtypes in islet finding, there is no colocalization of Kv2.1 and the gluca- cells and to obviate the possibility of contamination from gon signals (Fig. 2D). The presence of Kv2.1 protein in surrounding tissues, each of the most prominent islet Kv islets, but not in surrounding pancreatic tissue, was con- channel subtypes identified by PCR were tested by either firmed by immunostaining (Fig. 2E). Taken together, our or both ISH and IHC and compared with cell markers in results are consistent with the expression of Kv2.1 in sections of either Rhesus or human pancreatic tissue. pancreatic ␤-cells.

DIABETES, VOL. 53, MARCH 2004 603 Kv CHANNELS IN PRIMATE ISLETS

TABLE 3 Summary of expression of Kv channels in primate islets Subtype PCR ISH/IHC Subtype PCR ISH/IHC Kv1.1 Ϫ ND* Kv4.1 ϩ nonislet Kv1.2 Ϫ ND Kv4.3 Ϫ ND Kv1.3 ϩ ND Kv4.4 (ϩ)ND Kv1.4 Ϫ ND Kv5.1 Ϫ ND Kv1.5 Ϫ ND Kv6.1 ϩ␣-cell Kv1.6 ϩ nonislet Kv6.2 ϩ␤-cell Kv1.7 ϩ ND Kv6.3 Ϫ ND Kv2.1 ϩ␤-cell Kv8.1 Ϫ ND Kv2.2 ϩ␦-cell Kv9.1 (ϩ)ND Kv3.1 ϩ␣-cell Kv9.2 ϩ non–␤-cell Kv3.2 ϩ␤-cell Kv9.3 ϩ␤-cell Kv3.3 ϩ nonislet Kv10.1 ϩ ND Kv3.4 (ϩ)* ND Kv11.1 ϩ ND *Seen in one of three islet preparations.

Expression of Kv3.2 was studied with the same double- ␤-cells, we first performed an ISH experiment using Kv9.2- labeling protocol used for Kv2.1. Kv3.2 antisense oligo- specific oligonucleotide probes, but could not detect a nucleotides (Fig. 3G) and insulin antibody (Fig. 3H) significant signal over background levels (data not labeled the same population of islet cells (Fig. 3I), sug- shown). We then used a Kv9.2 riboprobe for ISH and the gesting that Kv3.2 is expressed in ␤-cells. insulin antibody for IHC experiments. The results in Fig. Because both Kv2.1 and Kv3.2 are expressed in ␤-cells, 5A–C show that the Kv9.2 mRNA is expressed in islets, but we tested for their colocalization in pancreatic sections. does not colocalize with insulin. Accordingly, we probed with a Kv3.2 antibody (Fig. 3J) The double-staining protocol with insulin antibody was and with Kv2.1 antisense oligonucleotides (Fig. 3K). As used to determine the cell distribution of Kv9.3 mRNA. expected, both positive cell populations overlap com- Kv9.3-positive cells completely overlap with insulin-con- pletely, further confirming that Kv2.1 and Kv3.2 are coex- taining cells (Fig. 5D–F), but not with those cells contain- pressed in ␤-cells, based on colocalization with insulin and ing glucagon (Fig. 5I). These data suggest that Kv9.3 each other (Fig. 3L). mRNA is exclusively expressed in ␤-cells. Although Kv10.1 Kv2.2 and Kv3.1 are found in ␦- and ␣-cells, respec- and Kv11.1 are found in human islets by RT-PCR (see tively, but not in ␤-cells. Pancreatic tissue was probed above), we did not attempt to further characterize the cell for Kv2.2 mRNA and insulin, using the same protocol as distribution of these two silent subunits. for Kv2.1 (Fig. 2F and G). In marked contrast to Kv2.1, Kv subunits not found in the islet. Other subunits, such colocalization of Kv2.2 with insulin was not observed (Fig. as Kv1.6, Kv3.3, and Kv4.1, were identified by RT-PCR in 2H). Subsequently, a section was probed with Kv2.2 anti- human islets. In ISH/IHC protocols, the three subtypes sense oligonucleotides and a somatostatin antibody. Both appear to be located outside of the islet, and no colocal- probes labeled the same cell population (Fig. 2I–K), ization with insulin was observed for any of them (Fig. indicating that Kv2.2 is expressed in the ␦-cells of the islet. 6A–I). Morphologically, the tubular configuration of the The double staining of Kv3.1 mRNA (Fig. 3A) and insulin Kv1.6-expressing cells suggests that these cells could be (Fig. 3B) did not overlap, suggesting that Kv3.1 is not acinar, Schwann, or nerve cells (41–43). expressed in ␤-cells (Fig. 3C). Subsequently, a section was The expression of Kv channels in islet cell types other probed with Kv3.1 antisense oligonucleotides (Fig. 3D) than ␣-, ␤-, or ␦-cells might contribute to the observed PCR and a glucagon antibody (Fig. 3E). Both probes labeled the signals (Fig. 1). However, this contribution does not same cell population (Fig. 3F), indicating that Kv3.1, in appear to be significant because, by ISH, Kv channels that contrast to Kv3.2, is expressed in pancreatic ␣-cells. coexpress with either insulin, somatostatin, or glucagon Distribution of “electrically silent” subunits in pan- are only found in cells that contain that marker. If other creatic sections. Several silent subunits were found to be cell types were responsible for the PCR signal, at least present in human islet cDNA by RT-PCR (see above). some cells would be expected to show the Kv signal Human pancreatic sections were probed for Kv6.1 and separate from the marker. Kv6.2 channels with labeled antisense oligonucleotides (Table 2) and with cell marker–specific antibodies in a double-staining experiment. The results shown in Fig. DISCUSSION 4A–H demonstrate that Kv6.1 is colocalized with glucagon, The identification of the molecular components for the IDR but not with insulin. In Figs. 4I–L, Kv6.2 appears to in human ␤-cells is critical for the development of an colocalize with all of the insulin-containing cells. However, inhibitor of this channel that would function as a glucose- in a number of islets, Kv6.2 labels only a major fraction of dependent insulin secretagogue for the treatment of type 2 the insulin-containing cells (data not shown). This distri- diabetes. This study is an attempt to identify the Kv bution could indicate the existence of a subset of ␤-cells subunits that are present in human ␤-cells, with the within some islets. ultimate goal of correlating their biophysical and pharma- To determine whether Kv9.2 is expressed in pancreatic cological properties with the currents found in ␤-cells. The

604 DIABETES, VOL. 53, MARCH 2004 L. YAN AND ASSOCIATES

FIG. 6. Kv1.6, Kv3.3, and Kv4.1 channels are not expressed in Rhesus pancreatic ␤-cells. In all sections, cell nuclei were stained with DAPI and are illustrated in blue. All images in this figure were taken from sections of Rhesus pancreas A: Kv1.6 immunoreactive cells, detected with Texas red (red), are located outside of the islet. B: Insulin immunoreactivity labels the same islet as seen in panel A and delineates the ␤-cells. C: Combination of IHC for Kv1.6 (red) and insulin (green) demonstrate signals on different cells. There is no coexpression as judged by the lack of yellow color. D: Kv3.3 mRNA-containing cells (red) in a section of pancreas. E: Insulin protein-containing cells (green) in the same islet. F: No colocalization (no yellow color) of Kv3.3 mRNA and insulin protein was detected. G: Kv4.1 mRNA-containing cells (red) in a section of pancreas. H: Insulin protein-containing cells (green) in the same section of the islet. I: No colocalization (no yellow color) of Kv4.1 mRNA and insulin protein was detected within the islet. initial ISH experiments on both human and Rhesus islets It is interesting that two silent subunits, Kv6.2 and demonstrate that expression of the Kv channels under Kv9.3, are also expressed in human ␤-cells. In heterolo- investigation exhibit the same pattern in both species. Our gous expression systems, these subunits are known to results combining RT-PCR with ISH and IHC strongly coassemble with subunits from either Kv2 or Kv3 families suggest that Kv2.1 and Kv3.2 are the major subunits in and to modify their function (24,47–49). It remains to be ␤-cells. In addition, silent subunits Kv6.2 and 9.3 are also determined if this also occurs in ␤-cells. In addition two present in ␤-cells (Table 3). other silent subunits, Kv10.1 and 11.1, previously reported In heterologous expression systems, Kv2.1 and Kv3.2 to be in pancreas and also known to associate with express delayed rectifier-type currents that resemble those subunits from Kv2 and Kv3 family (50), are found in islets. present in ␤-cells (44–46). Therefore, either one or the Their presence in ␤-cells and their significance will require ␤ other or both may contribute to IDR in -cells. Because the further investigation. tetraethylammonium ion and hanatoxin sensitivities of It is curious that members from two particular families individually heterologously expressed channels are quite distribute to different cell types within the islet. For different, it may be possible to use these tools to distin- instance, Kv2.1 and Kv3.2 distribute to the ␤-cell, whereas guish the relative contributions of these channel subtypes Kv2.2 and Kv3.1 are present in ␦- and ␣-cells, respectively. ␤ to the -cell IDR (44). This differential distribution has significant implications

DIABETES, VOL. 53, MARCH 2004 605 Kv CHANNELS IN PRIMATE ISLETS for the development of inhibitors that specifically target 14. Zunkler BJ, Trube G, Ohno-Shosaku T: Forskolin-induced block of delayed ϩ channels present in ␤-cells. rectifying K channels in pancreatic beta-cells is not mediated by cAMP. Pflugers Arch 411:613–619, 1988 There have been reports of Kv1 family channels in islets, 15. Philipson LH, Rosenberg MP, Kuznetsov A, Lancaster ME, Worley JF 3rd, ␤-cells, and insulin-secreting cell lines (16,51,52). In hu- Roe MW, Dukes ID: Delayed rectifier Kϩ channel overexpression in man islets, others (51) have reported that by RT-PCR transgenic islets and beta-cells associated with impaired glucose respon- Kv1.1, Kv1.2, and Kv1.4 are not found, while Kv1.5 and siveness. J Biol Chem 269:27787–27790, 1994 Kv1.6 were present. In our studies of human islets, only 16. Roe MW, Worley JF 3rd, Mittal AA, Kuznetsov A, DasGupta S, Mertz RJ, Kv1.3 and Kv1.6 were identified by RT-PCR, with a very Witherspoon SM 3rd, Blair N, Lancaster ME, McIntyre MS, Shehee WR, Dukes ID, Philipson LH: Expression and function of pancreatic beta-cell weak indication for Kv1.7. Kv1.6 is external to the islet, in delayed rectifier Kϩ channels: role in stimulus-secretion coupling. J Biol contrast to the recent report of its presence in rat ␤-cells Chem 271:32241–32246, 1996 (53). While Kv1.4 seems to be absent from human islets by 17. Eberhardson M, Tengholm A, Grapengiesser E: The role of plasma RT-PCR, it appears to be present in rat ␤-cells by Western membrane Kϩ and Ca2ϩ permeabilities for glucose induction of slow 2ϩ blot and PCR (51), but not in mouse by immunostaining Ca oscillations in pancreatic beta-cells. Biochim Biophys Acta 1283:67– 72, 1996 (53). These data could indicate differences in channel 18. Burge MR, Sood V, Sobhy TA, Rassam AG, Schade DS: Sulphonylurea- composition between species and highlights the impor- induced hypoglycaemia in type 2 diabetes mellitus: a review. Diabetes tance of the identification of the relevant subunits in Obes Metab 1:199–206, 1999 human ␤-cells. 19. Del Prato S, Aragona M, Coppelli A, Burge MR, Sood V, Sobhy TA, Rassam The identification of different channel types exclusive to AG, Schade DS: Sulfonylureas and hypoglycaemia: sulphonylurea-induced hypoglycaemia in type 2 diabetes mellitus: a review. Diabetes Nutr Metab each of the three major cell types found in the islet 15: 444–450, 2002 [discussion in 15:450–451, 2002] suggests that it may be possible to select for cell-type– 20. MacDonald PE, Salapatek AM, Wheeler MB: Glucagon-like peptide-1 dependent intervention through block of their respective receptor activation antagonizes voltage-dependent repolarizing Kϩ currents in ␤-cells: a possible glucose-dependent insulinotropic mechanism. IDRs. All of the Kv subunits tested in islets have been found in other tissues, but, in general, the exact combination of Diabetes 51 (Suppl. 3):S443–S447, 2002 ␤ 21. Kanno T, Gopel SO, Rorsman P, Wakui M: Cellular function in multicellular subunits in these tissues is unknown. For the -cell, system for hormone-secretion: electrophysiological aspect of studies on determination of IDR composition may aid significantly in alpha-, beta- and delta-cells of the pancreatic islet. Neurosci Res 42:79–90, the identification of a glucose-dependent insulin secreta- 2002 gogue applicable in type 2 diabetes, without the hypogly- 22. Coetzee WA, Amarillo Y, Chiu J, Chow A, Lau D, McCormack T, Moreno H, cemic liabilities found with K inhibitors. Nadal MS, Ozaita A, Pountney D, Saganich M, Vega-Saenz de Miera E, Rudy ATP B: Molecular diversity of Kϩ channels. Ann N Y Acad Sci 868:233–285, 1999 ACKNOWLEDGMENTS 23. Zhao B, Rassendren F, Kaang BK, Furukawa Y, Kubo T, Kandel ER: A new class of noninactivating Kϩ channels from aplysia capable of contributing We would like to thank Drs. Maria Garcia, Jim Herrington, to the resting potential and firing patterns of neurons. Neuron 13:1205– and Owen B. McManus for critical discussion and con- 1213, 1994 24. Stocker M, Hellwig M, Kerschensteiner D: Subunit assembly and domain structive comments. We would also like to thank Dr. H.G. ϩ Knaus for generously supplying antibodies for Kv1.6 and analysis of electrically silent K channel alpha-subunits of the rat Kv9 subfamily. J Neurochem 72:1725–1734, 1999 Dr. Jonathan Lakey for supplying human islets. 25. Zhu XR, Netzer R, Bohlke K, Liu Q, Pongs O: Structural and functional characterization of Kv6.2 a new gamma-subunit of voltage-gated potassium REFERENCES channel. Receptors Channels 6:337–350, 1999 1. Hille B: Ion channels of excitable membranes. In Potassium Channels and 26. Kerschensteiner D, Stocker M: Heteromeric assembly of Kv2.1 with Kv9.3: Chloride Channels. Sunderland, U.K., Sinauer Associates, 2001, p. 137–167 effect on the state dependence of inactivation. Biophys J 77:248–257, 1999 2. Dukes ID, Philipson LH: Kϩ channels: generating excitement in pancreatic 27. Sano Y, Mochizuki S, Miyake A, Kitada C, Inamura K, Yokoi H, Nozawa K, ␤-cells. Diabetes 45:845–853, 1996 Matsushime H, Furuichi K: Molecular cloning and characterization of 3. Aguilar-Bryan L, Bryan J: Molecular biology of adenosine triphosphate- Kv6.3, a novel modulatory subunit for voltage-gated Kϩ channel Kv2.1. sensitive potassium channels. Endocr Rev 20:101–135, 1999 FEBS Lett 512:230–234, 2002 4. Miki T, Nagashima K, Seino S: The structure and function of the ATP- 28. Su J, Yu H, Lenka N, Hescheler J, Ullrich S: The expression and regulation sensitive Kϩ channel in insulin-secreting pancreatic beta-cells. J Mol of depolarization-activated Kϩ channels in the insulin-secreting cell line Endocrinol 22:113–123, 1999 INS-1. Pflugers Arch 442:49–56, 2001 5. Ashcroft FM, Gribble FM: Correlating structure and function in ATP- 29. Betsholtz C, Baumann A, Kenna S, Ashcroft FM, Ashcroft SJ, Berggren PO, sensitive Kϩ channels. Trends Neurosci 21:288–294, 1998 Grupe A, Pongs O, Rorsman P, Sandblom J: Expression of voltage-gated 6. Yokoshiki H, Sunagawa M, Seki T, Sperelakis N: ATP-sensitive Kϩ chan- Kϩ channels in insulin-producing cells: analysis by polymerase chain nels in pancreatic, cardiac, and vascular smooth muscle cells. Am J reaction. FEBS Lett 263:121–126, 1990 Physiol 274:C25–C37, 1998 30. Philipson LH, Hice RE, Schaefer K, LaMendola J, Bell GI, Nelson DJ,

7. Ashcroft SJ: The beta-cell KATP channel. J Membr Biol 176:187–206, 2000 Steiner DF: Sequence and functional expression in Xenopus oocytes of a 8. Petit P, Loubatieres-Mariani MM: Potassium channels of the insulin- human insulinoma and islet potassium channel. Proc Natl Acad SciUSA secreting B cell. Fundam Clin Pharmacol 6:123–134, 1992 88:53–57, 1991

9. Kukuljan M, Goncalves AA, Atwater I: Charybdotoxin-sensitive KCa chan- 31. Bubacz DG, Dukes ID, McLean EW, Noe RA, Peat AJ, Szewczyk JR, nel is not involved in glucose-induced electrical activity in pancreatic Thomson SA, Worley JF: Kv2.1 antagonists. WO 99/32487, 1999 (PCT/ beta-cells. J Membr Biol 119:187–195, 1991 EP98/08085) 10. Gopel SO, Kanno T, Barg S, Eliasson L, Galvanovskis J, Renstrom E, 32. MacDonald PE, Sewing S, Wang J, Joseph JW, Smukler SR, Sakellaropou- Rorsman P: Activation of Ca2ϩ-dependent Kϩ channels contributes to los G, Saleh MC, Chan CB, Tsushima RG, Salapatek AM, Wheeler MB: rhythmic firing of action potentials in mouse pancreatic beta cells. J Gen Inhibition of Kv2.1 voltage-dependent Kϩ channels in pancreatic beta-cells Physiol 114:759–770, 1999 enhances glucose-dependent insulin secretion. J Biol Chem 277:44938– 11. Rorsman P, Eliasson L, Renstrom E, Gromada J, Barg S, Gopel S: The cell 44945, 2002 physiology of biphasic insulin secretion. News Physiol Sci 15:72–77, 2000 33. MacDonald PE, Wang G, Tsuk S, Dodo C, Kang Y, Tang L, Wheeler MB, 12. Smith PA, Bokvist K, Arkhammar P, Berggren PO, Rorsman P: Delayed Cattral MS, Lakey JR, Salapatek AM, Lotan I, Gaisano HY: Synaptosome- rectifying and calcium-activated Kϩ channels and their significance for associated protein of 25 kilodaltons modulates Kv2.1 voltage-dependent action potential repolarization in mouse pancreatic beta-cells. J Gen Kϩ channels in neuroendocrine islet beta-cells through an interaction with Physiol 95:1041–1059, 1990 the channel N terminus. Mol Endocrinol 16:2452–2461, 2002 13. Smith PA, Bokvist K, Rorsman P: Demonstration of A-currents in pancre- 34. Koch RO, Wanner SG, Koschak A, Hanner M, Schwarzer C, Kaczorowski atic islet cells. Pflugers Arch 413:441–443, 1989 GJ, Slaughter RS, Garcia ML, Knaus HG: Complex subunit assembly of

606 DIABETES, VOL. 53, MARCH 2004 L. YAN AND ASSOCIATES

neuronal voltage-gated Kϩ channels: basis for high-affinity toxin interac- 42. Ushiki T, Watanabe S: Distribution and ultrastructure of the autonomic tions and pharmacology. J Biol Chem 272:27577–27581, 1997 nerves in the mouse pancreas. Microsc Res Tech 37:399–406, 1997 35. Fiedor P, Rowinski W, Licinska I, Mazurek AP, Hardy MA: The survival 43. Castorina S, Romeo R, Marcello MF: Immunohistochemical study of identification of pancreatic islets of Langerhans: in vitro and in vivo effects intrinsic innervation in the human pancreas. Boll Soc Ital Biol Sper 72:1–7, of two dithizone preparations on staining of rat and human islets of 1996 Langerhans-preliminary study (Part I). Acta Pol Pharm 52:431–436, 1995 44. Swartz KJ, MacKinnon R: An inhibitor of the Kv2.1 potassium channel 36. Lynch KR, O’Neill GP, Liu Q, Im DS, Sawyer N, Metters KM, Coulombe N, isolated from the venom of a Chilean tarantula. Neuron 15:941–949, 1995 Abramovitz M, Figueroa DJ, Zeng Z, Connolly BM, Bai C, Austin CP, 45. Rudy B, McBain CJ: Kv3 channels: voltage-gated Kϩ channels designed for Chateauneuf A, Stocco R, Greig GM, Kargman S, Hooks SB, Hosfield E, high-frequency repetitive firing. Trends Neurosci 24:517–526, 2001 Williams DL, Jr, Ford-Hutchinson AW, Caskey CT, Evans JF: Character- 46. Kelly RP, Sutton R, Ashcroft FM: Voltage-activated calcium and potassium ization of the human cysteinyl leukotriene CysLT1 receptor. Nature currents in human pancreatic beta-cells. J Physiol 443:175–192, 1991 399:789–793, 1999 47. Post MA, Kirsch GE, Brown AM: Kv2.1 and electrically silent Kv6.1 37. Figueroa DJ, Hess JF, Ky B, Brown SD, Sandig V, Hermanowski-Vosatka A, potassium channel subunits combine and express a novel current. FEBS Twells RC, Todd JA, Austin CP: Expression of the type I diabetes- Lett 399:177–182, 1996 associated gene LRP5 in macrophages, vitamin A system cells, and the 48. Salinas M, Duprat F, Heurteaux C, Hugnot JP, Lazdunski M: New modu- islets of Langerhans suggests multiple potential roles in diabetes. J His- latory alpha subunits for mammalian Shab Kϩ channels. J Biol Chem tochem Cytochem 48:1357–1368, 2000 272:24371–24379, 1997 38. Kalman K, Nguyen A, Tseng-Crank J, Dukes ID, Chandy G, Hustad CM, 49. Blaine JT, Ribera AB: Heteromultimeric potassium channels formed by Copeland NG, Jenkins NA, Mohrenweiser H, Brandriff B, Cahalan M, members of the Kv2 subfamily. J Neurosci 18:9585–9593, 1998 Gutman GA, Chandy KG: Genomic organization, chromosomal localiza- 50. Ottschytsch N, Raes A, Van Hoorick D, Snyders DJ: Obligatory heterotet- tion, tissue distribution, and biophysical characterization of a novel ramerization of three previously uncharacterized Kv channel alpha-sub- mammalian Shaker-related voltage-gated potassium channel, Kv1.7. J Biol units identified in the human genome. Proc Natl Acad SciUSA Chem 273:5851–5857, 1998 99:7986–7991, 2002 39. Klein T, Ling Z, Heimberg H, Madsen OD, Heller RS, Serup P: Nestin is 51. MacDonald PE, Ha XF, Wang J, Smukler SR, Sun AM, Gaisano HY, expressed in vascular endothelial cells in the adult human pancreas. Salapatek AM, Backx PH, Wheeler MB: Members of the Kv1 and Kv2 J Histochem Cytochem 51:697–706, 2003 voltage-dependent Kϩ channel families regulate insulin secretion. Mol 40. Baetens D, Vasko M, Unger RH, Orci L: Ultrastructural detection of Endocrinol 15:1423–1435, 2001 granulated cells in the autonomic ganglia of the rat pancreas. Diabetologia 52. MacDonald PE, Wheeler MB: Voltage-dependent Kϩ channels in pancreatic 28:841–846, 1985 beta cells: role, regulation and potential as therapeutic targets. Diabetolo- 41. Persson-Sjogren S, Zashihin A, Forsgren S, Ushiki T, Watanabe S, Casto- gia 46:1046–1062, 2003 rina S, Romeo R, Marcello MF: Nerve cells associated with the endocrine 53. Gopel SO, Kanno T, Barg S, Rorsman P: Patch-clamp characterisation of pancreas in young mice: an ultrastructural analysis of the neuroinsular somatostatin-secreting-cells in intact mouse pancreatic islets. J Physiol complex type I. Histochem J 33:373–378, 2001 528:497–507, 2000

DIABETES, VOL. 53, MARCH 2004 607