Heterozygous missense mutations in the pore-forming subunit Kir6.2 of ATP-sensitive K+ channels (KATP channels) have recently been shown to cause permanent neonatal diabetes mellitus (PNDM). Functional studies demonstrated that PNDM mutations reduce KATP channel sensitivity to ATP inhibition, resulting in gain of channel function. However, the impact of these mutations on channel expression has not been examined. Here, we show that PNDM mutations, including Q52R, V59G, V59M, R201C, R201H, and I296L, not only reduce channel ATP sensitivity but also impair channel expression at the cell surface to varying degrees. By tagging the PNDM Kir6.2 mutant V59G or R201H with an additional mutation, N160D, that confers voltage-dependent polyamine block of KATP channels, we demonstrate that in simulated heterozygous state, all surface channels are either wild-type or heteromeric channels containing both wild-type and mutant Kir6.2 subunits. Comparison of the various PNDM mutations in their effects on channel nucleotide sensitivity and expression, as well as disease phenotype, suggests that both channel-gating defect and expression level may play a role in determining disease severity. Interestingly, sulfonylureas significantly increase surface expression of certain PNDM mutants, suggesting that the efficacy of sulfonylurea therapy may be compromised by the effect of these drugs on channel expression.
Pancreatic ATP-sensitive K+ channels (KATP channels), each consisting of four pore-forming Kir6.2 subunits and four regulatory sulfonylurea receptor one (SUR1) subunits, link β-cell metabolism to insulin secretion (1–3). The activity of KATP channels is governed mainly by the dynamics of intracellular adenine nucleotides ATP and ADP at the channel site during glucose metabolism (1,4). Both nucleotides can stimulate or inhibit channel activity depending on their relative concentrations and whether Mg2+ is present. Inhibition of channels by nucleotides is mediated by the Kir6.2 subunit and does not require Mg2+ (5,6), whereas nucleotide stimulation is conferred by the SUR1 subunit and requires Mg2+ (7,8). The physiological activity of KATP channels in β-cells is thus a balance between nucleotide inhibition and nucleotide stimulation. During glucose stimulation, ATP concentrations increase and ADP concentrations decrease, resulting in KATP channel closure. Because KATP channels carry the dominant conductance in high-input resistance β-cells at resting state, closure of KATP channels leads to membrane depolarization, which in turn leads to opening of voltage-gated calcium channels, calcium influx, and insulin release.
Recent studies have established heterozygous missense mutations in Kir6.2 as a major cause underlying permanent neonatal diabetes mellitus (PNDM) (9–16). As Kir6.2 is also a constituent of KATP channel subtypes expressed outside of pancreas, including cardiac muscle, skeletal muscle, and brain, some mutations have been reported to cause muscle weakness, dysmorphic features, developmental delay, epilepsy, in addition to neonatal diabetes (referred to as DEND syndrome) (15,17). Functional studies of mutant channels have revealed reduced ATP sensitivity as a common defect (11,18,19). While reduced channel sensitivity to ATP provides an explanation for how these mutations lead to KATP channel overactivity in β-cells at high glucose, and thereby diabetes (20,21), some puzzles remain. For example, in some mutations, such as R201C and R201H, reduced ATP sensitivity was clearly observed in homomeric mutant channels; however, no significant difference in ATP sensitivity from wild-type channels could be detected under the condition that simulates heterozygous expression in patients (11,18,21). A potential explanation for the lack of detectable difference could be that the mutant is not expressed as efficiently as the wild-type Kir6.2, a hypothesis that has not been tested. Furthermore, the extent of ATP sensitivity reduction seen in the different mutant channels does not always match well with disease severity. For example, the mutation V59M results in channels much more sensitive to ATP than R201C or R201H, yet it is associated with more severe disease phenotype. These observations suggest additional factors might contribute to the pathogenic potency of a mutation.
Here, we examined the effects of several PNDM mutations, including Q52R, V59G, V59M, R201C, R201H, and I296L, on KATP channel expression. We found that all of them lead to reduced surface expression, to varying degrees, of KATP channels reconstituted in mammalian cells. By tagging the V59G or R201H mutant Kir6.2 with the N160D inward rectification mutation, which confers voltage-dependent spermine block of KATP channels (22), we demonstrate that in simulated heterozygous expression condition, nearly all channels present at the cell surface are either pure wild-type or heteromeric channels containing a mixture of wild-type and mutant Kir6.2. We present evidence that the expression level of a mutation plays a role in determining the extent of β-cell dysfunction. Moreover, we show that sulfonylureas significantly enhance surface expression of some PNDM mutant channels, suggesting the efficacy of sulfonylurea therapy may be compromised by its effect on channel expression.
RESEARCH DESIGN AND METHODS
Rat Kir6.2 cDNA is in pCDNAI vector and SUR1 in pECE. Site-directed mutagenesis was carried out using the QuickChange site-directed mutagenesis kit (Strategene) and the mutation confirmed by sequencing. Construction of adenovirus carrying R201H Kir6.2 cDNA is as described previously (23).
Cells grown in 35-mm dishes were transfected using FuGene6 (Roche Applied Science, Indianapolis, IN) with 0.4 μg rat Kir6.2 and 0.6 μg of a SUR1 tagged with a FLAG-epitope (DYKDDDDK) at the NH2-terminus (referred to as fSUR1) (24). Cells were lysed 48–72 h post-transfection in 20 mmol/l HEPES (pH 7.0, 5 mmol/l EDTA, 150 mmol/l NaCl, 1% Nonidet P-40 [Igapel]) with Complete protease inhibitors (Roche Applied Science) and lysate analyzed by standard Western blot procedures (24). Mouse monoclonal antibodies for FLAG (M2) and α-tubulin were from Sigma (St. Louis, MO), rabbit polyclonal antibodies for Kir6.2 and mouse monoclonal antibody for CIEBP homolog protein (CHOP) from Santa Cruz Biotechnology (Santa Cruz, CA), and mouse monoclonal antibody for BiP from BD Transduction Laboratories (San Diego, CA).
COSm6 cells transiently expressing fSUR1 and Kir6.2 were fixed with 2% paraformaldehyde 48–72 h posttransfection, incubated with anti-FLAG antibody (10 μg/ml) followed by horseradish peroxidase–conjugated anti-mouse secondary antibodies (1:1,000 dilution; Amersham), and SuperSignal ELISA Femto luminol solution (24,25). Chemiluminescence signal was read in a TD-20/20 luminometer (Turner Designs). Results of each experiment are the average of 2–3 dishes, and each data point shown in figures is the average of 3–5 independent experiments.
INS-1 cells clone 832/13 (26) were plated in 24-well plates and cultured for 24 h in RPMI-1640 with 11.1 mmol/l d-glucose (Invitrogen) supplemented with 10% fetal bovine serum, 100 units/ml penicillin, 100 μg/ml streptomycin, 10 mmol/l HEPES, 2 mmol/l glutamine, 1 mmol/l sodium pyruvate, and 50 μmol/l β-mercaptoethanol. Recombinant adenoviruses containing wild-type or R201H Kir6.2 with desired titers were then used to infect cells as described previously (23).
Insulin secretion assay.
Insulin secretion assays in INS-1 cells were performed as described previously (23). Twenty-four hours post–virus infection, the culture medium was replaced by RPMI-1640 with 5 mmol/l glucose and cells incubated for 18 h. Insulin secretion was assayed by preincubating cells in Hanks’ balanced salt solution (in mmol/l: 114 NaCl, 4.7 KCl, 1 MgCl2, 1.2 KH2PO4, 1.16 MgSO4, 20 HEPES, 2.5 CaCl2, 25.5 NaHCO3, and 0.2% BSA; pH ∼7.2) (26) containing 3 mmol/l glucose for 2 h before stimulation with Hanks’ balanced salt solution containing 3 or 12 mmol/l glucose for 2 h. Insulin content was determined using Immunochem coated-tube insulin radioimmunoassay from ICN Pharmaceuticals (Costa Mesa, CA). Insulin release at 12 mmol/l glucose was normalized to that at 3 mmol/l glucose and expressed as fold increase.
For experiments in Figs. 3 and 4 and the online appendix (availabe at http://diabetes.diabetesjournal.org), inside-out patch-clamp recording was performed in COSm6 cells expressing wild-type or mutant KATP channels. Recording pipettes had average resistance of ∼1.0–1.5 MΩ. All recordings were made with the Axopatch 1D amplifier and Clampex 8.1 (Axon, Foster City, CA) at room temperature −50 mV with symmetrical K-INT solution containing (in mmol/l) 140KCl, 10 K-HEPES, and 1 KEGTA (pH 7.2 with KOH). For measuring ATP sensitivity, 1 mmol/l EDTA was included in K-INT to avoid channel rundown (27). For measuring MgATP sensitivity, MgCl2 (at concentration equal to that of ATP) was added to K-INT as Mg2+ source. For spermine block experiments, 20 μmol/l spermine (Sigma) was added to K-INT. Excised patches were subjected to voltage ramps (200 ms) between +100 and −100 mV, with each ramp preceded by 20 ms holding potential at +140 mV to saturate channels with spermine (28). Currents obtained in the presence of 10 mmol/l ATP were taken as leakage and subtracted from the total currents for analysis. Only patches with currents >1 nA (at −50 mV) were included for analysis. Whole-cell patch-clamp recording was used to measure INS-1 cell resting membrane potential (23). Cells were preincubated in 12 mmol/l glucose for 3 h before recording. During recording, cells were bathed in Tyrode’s solution consisting of (in mmol/l) 137 NaCl, 5.4 KCl, 1.8 CaCl2, 0.5 MgCl2, 5 HEPES, 3 NaHCO3, and 0.16 NaH2PO4, with 12 mmol/l glucose. Pipette solution contained (in mmol/l) 10 KCl, 130 Kgluconate, 10 HEPES, 1 EGTA, 3 MgCl2, and 5 ATP.
Data fitting was performed with Origin 6.1. For estimating expression ratio of wild-type and N160D-tagged mutant subunit, the Grel/V curve obtained in the presence of spermine was fitted as described previously (28) with the sum of five individual Boltzmann equations:
where Ai, Vi, and zi are the amplitude, voltage of half-maximal inhibition, and effective valency, respectively, of the ith component (with 0, 1, 2, 3, or 4 mutant subunits). We assumed that the fitted amiplitude (Ai) corresponds to the probability (px) of formation of each channel component, with wild-type and mutant Kir6.2 subunits having equal probability of being incorporated into the channel complex, following the binomial distribution:
where n is the number of Kir6.2 subunits in a channel, x is the number, 0 to n, of wild-type subunits in a given channel component, P is the probability of inclusion of a wild-type subunit, and (1 − P) is the probability of inclusion of a mutant subunit. The P value that gave the best fit was taken as the fraction of wild-type subunits present in the channel population in a given patch.
Effects of PNDM mutations on KATP channel expression.
To study KATP channel expression, we chose the following Kir6.2 mutations: Q52R, V59G, V59M, R201C, R201H, and I296L to include both the mild PNDM disease phenotype and the more severe DEND phenotype. Although functional data reported to date has been largely obtained using the Xenopus oocyte expression system (11,18), this system is less suitable for studying channel maturation and trafficking since many misfolded proteins that fail to mature in mammalian cells are tolerated in Xenopus oocytes (29,30). We therefore examined channel expression in the mammalian cell line COSm6. Western blot analysis was used to assess steady-state channel protein expression level in cells cotransfected with Kir6.2 and fSUR1. Channel assembly occurs in the endoplasmic reticulum. As the correctly assembled channel complex exits the endoplasmic reticulum and travels through the Golgi, the two N-linked glycosylation sites in SUR1 are further modified, giving rise to the complex-glycosylated form that migrates slower on the SDS gel (the upper band) than the core-glycosylated form (the lower band). The intensity of the upper fSUR1 band is thus indicative of the level of assembled channel complexes that have trafficked beyond the medial Golgi. As shown in Fig. 1A, many of the mutants exhibited reduced upper fSUR1 band, indicating reduced channel expression. In some mutants (for example, R201H) a reduced steady-state Kir6.2 protein level was also observed, presumably due to decreased synthesis or increased degradation; the reduced Kir6.2 protein level would lead to reduced channel expression. To ensure that the reduced expression is not unique to COSm6 cells, we next examined mutant channel expression in HEK293 cells and a neuroblastoma cell line N2a. Figure 1B and C show that the expression of the R201H mutant in these two cell lines was similarly reduced (other mutants also exhibited reduced expression as seen in COSm6 cells, not shown). Surface expression was further quantified using a chemiluminescence-based immunoassay (25,31). Consistent with Western blot results, mutant channels expressed in COS cells exhibited reduced surface expression to varying degrees, with Q52R, V59G, R201C, R201H, and I296L reduced by ∼80% and V59M moderately reduced by ∼50% (Fig. 1D).
One possible mechanism of reduced channel surface expression is that the mutant Kir6.2 proteins are unable to fold or assemble correctly. If such protein is not effectively degraded and accumulates in the endoplasmic reticulum, it may elicit unfolded protein response (UPR) and may even trigger apoptosis if the endoplasmic reticulum stress cannot be overcome (32). Although no apparent accumulation of Kir6.2 was seen for our mutants, we decided to test the UPR scenario directly since mutant protein-induced UPR and cell apoptosis could be a contributing mechanism to neonatal diabetes. We examined two marker proteins, BiP (GRP78), an endoplasmic reticulum chaperone protein, and CHOP (GADD153), an endoplasmic reticulum stress-associated apoptosis transcription factor, whose levels are known to be upregulated during UPR and UPR-induced apoptosis (33). No difference in either BiP or CHOP protein level was detected between control COSm6 cells, cells expressing wild-type KATP channels and cells expressing PNDM mutant channels (Fig. 2). In contrast, cells treated with tunicamycin, which blocks N-linked protein glycosylation and induces UPR, exhibited increased level of both BiP and CHOP (34). These results indicate that the mutant Kir6.2 proteins do not cause significant endoplasmic reticulum stress and apoptosis, at least in COSm6 cells, within the time course of our experiments.
Assessing the relative abundance of wild-type and mutant Kir6.2 in surface channels in simulated heterozygous expression by N160D mutation-dependent and voltage-dependent spermine block.
The reduced surface expression of the mutant channels observed above suggests that under heterozygous expression condition as seen in patients few surface channels would be pure (homomeric) mutant channels. To test this, we monitored the relative number of surface channels containing 0, 1, 2, 3, or all 4 mutant subunits in heterozygous expression by tagging the mutant Kir6.2 subunits with the N160D mutation, which confers strong inward rectification in the presence of spermine (22,28). Previous studies have shown that the extent of rectification depends on the number of Kir6.2 subunits carrying the N160D mutation and that the relative conductance-voltage (Grel-V) curve of each channel population containing 0, 1, 2, 3, or 4 N160D subunits can be fitted by a single Boltzmann function (22). In simulated heterozygous expression, the overall Grel-V curve will therefore be the sum of the five individual Boltzmann functions with the relative conductance amplitude of each Boltzmann function corresponding to the relative abundance of each channel population (28). Assuming wild-type and mutant Kir6.2 protein have equal probability of being incorporated into the channel complex, the distribution of the five-channel populations should be binomial with the relative abundance of each channel population dictated by the relative amount of the wild-type versus mutant Kir6.2 proteins. This approach can thus be used to estimate the ratio of wild-type to mutant Kir6.2 in a single membrane patch. In cells expressing wild-type and R201H/N160D Kir6.2 at 1:1 cDNA ratio, the patches consistently contained ∼83% of wild-type protein and ∼17% of R201H/N160D protein (n = 7), in good agreement with biochemical data that R201H Kir6.2 is not equally expressed as wild-type Kir6.2 (see Fig. 1). Control experiments confirmed that the N160D mutation does not affect expression of the R201H channels (not shown). Decreasing cDNA ratio of wild-type to mutant Kir6.2 to 1:2 or 1:4 resulted in concordant change in the estimated protein ratio (wild-type–to–mutant ratio of 0.72:0.28 for cDNA ratio of 1:2, n = 6, and 0.60:0.40 for cDNA ratio of 1:4, n = 7), as evidenced by the more severe inward-rectification in intracellular 20 μmol/l spermine (an example of 1:4 cDNA ratio expression experiment is shown in Fig. 3). Thus, in simulated heterozygous state, although the cDNA ratio of wild type to mutant is 1:1, the protein ratio is ∼1:0.2. These results led us to conclude that in simulated heterozygous expression, most surface channels are either pure wild type or contain a mixture of wild-type and R201H mutant Kir6.2 subunits, with the population of homR201H mutant channels being <0.1% [(0.17)4]. Examination of cells coexpressing wild-type and V59G/N160D Kir6.2 at 1:1 cDNA ratio in the spermine block experiments similar to those described for R201H also confirmed reduced expression of the V59G mutant in the surface membrane (wild-type–to–V59G/N160D protein ratio of 0.83:0.17, n = 7). The above functional data further corroborate the results from Western blots and chemiluminescence assays that many PNDM mutations render reduced surface expression of KATP channels.
Heterozygous expression of R201H results in channels that are overactive in physiological concentrations of MgATP.
If only <0.1% of homR201H mutant channels are expressed at the cell surface, how could heterozygous R201H mutation increase the physiological activity of β-cell KATP channels? We reasoned that in cells, most ATP is complexed with Mg2+ (35). In the presence of Mg2+, ATP is hydrolyzed by SUR1; this leads to channel activation and effectively reduces the sensitivity of Kir6.2 to ATP inhibition (1,4,8). We therefore tested whether heterozygous R201H mutation might significantly affect channel sensitivity to MgATP. Indeed, we found that in simulated heterozygous expression (cDNA ratio of wild type to mutant = 1:1, i.e., protein ratio of ∼0.83:0.17 as shown in Fig. 3B), the resulting channel population exhibits significantly increased channel activity in mmol/l range of MgATP compared with pure wild-type channels (Fig. 4A). These observations are consistent with those recently reported by Gloyn et al. (36) and Proks et al. (37). Since the number of pure R201H mutant channels is close to 0 in heterozygous expression, based on the conservative estimation that the number of surface KATP channels in a single human β-cell is <800 (38,39), we conclude that the active channels in mmol/l MgATP concentration range are those containing 1, 2, or 3 R201H mutant Kir6.2 subunits (Fig. 4B).
Effect of R201H-Kir6.2 expression level on INS-1 cell membrane potential and insulin secretion response.
A question that arises from our observation of the varied expression efficiency of PNDM mutants (Fig. 1D) is whether the expression level plays a role in determining the extent of β-cell dysfunction. To address this issue, we expressed the R201H-Kir6.2 mutant in the insulin-secreting cell line INS-1 by infection with a recombinant adenovirus containing R201H-Kir6.2 cDNA. We chose two virus titers that gave one- or threefold excess of mutant protein expression compared with endogenous wild-type Kir6.2 protein (Fig. 5A). To ensure that cells expressing the mutant protein were not under severe endoplasmic reticulum stress or undergoing apoptosis, which could reduce insulin secretion, we again checked the level of BiP and CHOP. Similar to what we found in COS cells (Fig. 2), no change in either BiP or CHOP was detected during the time course of our experiments (Fig. 5B). Next, membrane potential and insulin secretion response to glucose stimulation (12 mmol/l) were measured by whole-cell patch-clamp recording and insulin radioimmunoassay. Figure 5C shows that the membrane potential of cells infected with R201H-Kir6.2 at 12 mmol/l glucose was more hyperpolarized, in an expression level–dependent manner, than uninfected controls cells or cells expressing similar levels of exogenous wild-type Kir6.2. Moreover, cells infected with the R201H-Kir6.2 virus had reduced insulin secretion response to 12 mmol/l glucose stimulation, again in an expression level–dependent manner (Fig. 5D). These results demonstrate directly the causal role of the R201H mutation in impairing glucose-stimulated insulin secretion. In addition, they provide evidence that the expression level of a PNDM mutant Kir6.2 indeed affects the extent of β-cell dysfunction.
Effects of sulfonylureas on surface expression of PNDM mutant channels.
Sulfonylurea therapy has been proposed as an alternative to insulin injection in PNDM patients with Kir6.2 mutations. We have previously shown that sulfonylureas increase surface expression of mutant channels with reduced surface expression due to SUR1 mutations associated with congenital hyperinsulinism (24,25). We asked whether sulfonylureas might also affect the expression efficiency of PNDM mutant channels. Interestingly, surface expression of several mutants in COSm6 cells was markedly increased when cells were treated with glibenclamide (1 μmol/l for 24 h; Fig. 6). The degree of increase appeared more pronounced in mutants with more severe expression defect. These results reveal the complex effects of sulfonylureas on not only gating but also expression of PNDM mutants.
In this study, we used both biochemical and electrophysiological approaches to demonstrate that many PNDM-associated Kir6.2 mutations reduce KATP channel surface expression, in addition to their well-documented effects on reducing channel ATP sensitivity (11–13,18,36,40). Detailed analyses of the R201H mutation show that in simulated heterozygous state the fraction of mutant Kir6.2 present in the plasma membrane is only ∼17% of the total surface Kir6.2 protein pool. The findings suggest that in patients, the number of homomeric mutant channels in a β-cell is almost zero and overactivity of KATP channels is therefore likely attributed to heteromeric channels containing both wild-type and mutant Kir6.2 subunits (Fig. 4). Interestingly, earlier studies showed that while reduced ATP sensitivity was obvious in homomeric R201H mutant channels, no significant change in ATP sensitivity was detectable in channels from simulated heterozygous expression condition. Our observation that R201H homomeric channels are poorly expressed (or nonexistent) at the cell surface may help explain why little change in channel ATP sensitivity was observed under heterozygous expression (11,18,21). The R201 residue is thought to be involved in ATP binding (40,41). Markworth et al. (42) have reported that only one of the four ATP binding sites in a Kir6.2 tetramer is necessary for ATP-induced channel closure; the wild-type subunit in heteromeric channels may thus be sufficient to confer wild-type–like ATP sensitivity. This interpretation, however, may not apply to other mutations, such as Q52R and V59G, that render the channel insensitive to ATP by increasing the intrinsic channel open probability rather than affecting ATP binding (18). In these mutants, it is possible that even one mutant subunit in the tetramer can increase channel open probability and indirectly reduce ATP sensitivity. Although heterozygous expression of R201H does not lead to significant reduction in ATP sensitivity, it does lead to significantly higher channel activity in physiological concentration of MgATP (Fig. 4A), consistent with that reported recently by others (36,37). Our analyses of other mutants also showed more pronounced channel overactivity in MgATP than in ATP (see online appendix). Collectively, these results lead us to conclude that for mutations presented in this study, the heteromeric channel population containing both wild-type and mutant Kir6.2 subunits plays a dominant role (as opposed to homomeric mutant channels) in causing disease by being more active in physiological concentration of MgATP.
Channel expression and disease.
Reduced surface expression of KATP channels due to mutations in the channel genes has been recognized as a significant mechanism contributing to loss of channel function in congenital hyperinsulinism (3,43–45). Somewhat unexpectedly, all of the PNDM mutations examined here also reduce channel surface expression to varying degrees. While intuitively one might predict that reduced surface expression should lead to loss of channel function, in the case of PNDM mutations the effect on expression is overruled by the effect of the mutations on gating (increased channel open probability in MgATP), leading to an overall “gain-of- channel-function” phenotype. This is not surprising considering that the activity of KATP channels in cells depends on both the number of channels in the plasma membrane and the open probability of channels under the physiological environment. Our finding also raises the question of whether the expression level of a PNDM mutation plays a role in its clinical manifestation. In this regard, our study examining the effect of R201H-Kir6.2 expression in INS-1 cells demonstrates that indeed the extent of β-cell dysfunction, as assessed by the ability of the cell to depolarize and secrete insulin upon glucose stimulation, is correlated with the expression level of the mutant. Among the PNDM mutations we examined, V59M stands out as having the highest expression level (Fig. 1). It is interesting to note that V59M channels exhibit the least change in ATP and MgATP sensitivities (Figs. 1 and 2; Tables 1 and 2, online appendix), yet it is associated with intermediate DEND that is more severe than the PNDM phenotype seen in another mutation, R201H, which causes much more reduced ATP sensitivity. We recognize that many factors are likely involved in determining the clinical phenotypes in patients; for instance, the R201C mutation has been reported to cause PNDM in some individuals but DEND in others (11,12,14). Nevertheless, it is tempting to speculate that the more severe phenotype seen in the V59M patients might be due in part to the mutant’s higher expression level (Table 2, online appendix).
Implications for sulfonylurea therapy.
The discovery that PNDM could be caused by overactive KATP channels has led to the proposal that sulfonylureas might be used as an alternative therapy to insulin injection in patients carrying Kir6.2 mutations (11). Indeed, several studies have reported successful treatment of patients with mutations that cause PNDM alone using sulfonylureas (11,13,46,47), although the long-term effectiveness of such treatment remains to be determined. We have previously shown that sulfonylureas markedly increase surface channel expression in two congenital hyperinsulinism-causing SUR1 mutations (25). In this study, we found that sulfonylureas also markedly increase surface expression level of several PNDM-associated mutant channels. Thus, on the one hand, sulfonylureas might reduce channel activity via their effects on channel gating; on the other, they might have the adverse effect of increasing channel activity by increasing mutant channel expression. Our finding, together with previous studies showing that some PNDM Kir6.2 mutations render the channel less sensitive to sulfonylurea inhibition (18,48), point to the potential complexity of sulfonylurea therapy. Future studies examining the effect of sulfonylureas on INS-1 cells expressing mutant channels may help evaluate the effectiveness of sulfonylurea therapy in specific mutations.
C.-W. Lin is currently affiliated with the Department of Brain and Cognitive Sciences, Massachusetts Institute of Technology, Cambridge, Massachussets.
Additional information for this article can be found in an online appendix at http://diabetes.diabetesjournals.org.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by National Institutes of Health Grant DK57699 (to S.-L. S.) and a Predoctoral Fellowship from the American Heart Association (to Y.-W. L.).
We thank Dr. Carol A. Vandenberg for providing the rat Kir6.2 clone and Dr. Chris Newgard for INS-1 cells clone 832/13.