Congenital Hyperinsulinism–Associated ABCC8 Mutations That Cause Defective Trafficking of ATP-Sensitive K⁺ Channels: Identification and Rescue

Peripheral blood was obtained from patients for isolation of genomic DNA. Individual exons and adjacent intron-exon boundaries of the ABCC8 gene were amplified by the PCR and screened for mutations by direct nucleotide sequencing as previously described (23). SUR1 cDNA and protein sequences were numbered according to Nestorowicz et al. (24), with nucleotide numbering beginning with the first methionine and including the alternatively spliced exon 17 sequence (NCBI accession no. L78224). The study protocol was reviewed and approved by the Institutional Review Board of The Children's Hospital of Philadelphia, and written informed consent was obtained from the parents of the patients.

FLAG epitope (DYKDDDDK) was inserted at the NH₂- terminus of the hamster SUR1 cDNA in the pECE plasmid as previously described (12). SUR1 point mutations were made using the QuickChange site-directed mutagenesis kit (Stratagene) and confirmed by DNA sequencing. Rat Kir6.2 cDNA (gift from Dr. Carol Vandenberg) is in pCDNAI vector. Multiple PCR clones were analyzed to avoid false results caused by undesired PCR-introduced mutations.

COSm6 cells were plated in six-well tissue culture plates and transfected with 0.6 μg fSUR1 and 0.4 μg rat Kir6.2 per dish using FuGene6 (Roche Applied Science). Cells were lysed in 20 mmol/l HEPES, pH 7.0, 5 mmol/l EDTA, 150 mmol/l NaCl, and 1% Nonidet P-40 with Complete protease inhibitors (Roche Applied Science) 48–72 h after transfection. Proteins were separated by SDS-PAGE (8%), transferred to nitrocellulose, analyzed by M2 anti-FLAG antibody (Sigma) followed by horseradish peroxidase (HRP)–conjugated anti-mouse secondary antibodies (GE Healthcare), and visualized by chemiluminescence (Super Signal West Femto; Pierce). Experimental procedures for the immunoblot shown in Fig. 5D using INS-1 cells are described in detail in the online appendix (available at http://dx.doi.org/10.2337/db07-0150).

COSm6 cells transfected with fSUR1 and Kir6.2 were plated onto coverslips for staining 48–72 h after transfection. For surface staining, living cells were incubated with anti-FLAG M2 mouse monoclonal antibody (10 μg/ml OptiMEM containing 0.1% BSA) for 1 h at 4°C, washed with ice-cold PBS, fixed with cold (−20°C) methanol for 10 min, and incubated with Cy-3–conjugated donkey anti-mouse secondary antibodies (Jackson) for 30 min at room temperature. After three 10-min washes in PBS/0.1% BSA and one 10-min wash in PBS, cells were viewed with an Olympus Fluoview confocal microscope. For total cellular staining, the same procedure was followed except the cells were fixed with cold (−20°C) methanol for 10 min before antibody incubation.

Cells were fixed with 2% paraformaldehyde for 30 min at 4°C, preblocked in PBS/0.1% BSA for 30 min, incubated in M2 anti-FLAG antibody (10 μg/ml) for 1 h, washed four times for 30 min in PBS/0.1% BSA, incubated in HRP-conjugated anti-mouse (1:1,000 dilution; Jackson) for 20 min, and washed again four times for 30 min in PBS/0.1% BSA. Chemiluminescence was quantified using a TD-20/20 luminometer (Turner Designs) after 5-s incubation in Power Signal Elisa Femto luminol solution (Pierce). All steps after fixation were performed at room temperature. Results of each experiment are the average of two dishes, and each data point shown in figures is the average of three to six experiments.

Cells were incubated overnight in culture medium containing ⁸⁶RbCl (1 μCi/ml) 48 h after transfection with fSUR1 and Kir6.2. Before measurement of ⁸⁶Rb⁺ efflux, the cells were incubated for 30 min at 25°C in Krebs-Ringer solution with metabolic inhibitors (2.5 μg/ml oligomycin plus 1 mmol/l 2-deoxy-d-glucose). At selected time points, the solution was aspirated from the cells and replaced with fresh solution, and at the end of a 40-min period, cells were lysed. The ⁸⁶Rb⁺ in the aspirated solution and the cell lysate was counted. The percentage of efflux at each time point was calculated as the cumulative counts in the aspirated solution divided by the total counts from the solutions and the cell lysates.

Inside-out patch clamp recordings were carried out as previously described using an Axopatch 1D amplifier (Axon, Foster City, CA) (16). Micropipettes were pulled from nonheparinized Kimble glass (Fisher Scientific) on a horizontal puller (Sutter Instrument, Novato, CA) with ∼1.5 MΩ electrode resistance. The bath and pipette solution (K-INT) had the following composition: 140 mmol/l KCl, 10 mmol/l K-HEPES, and 1 mmol/l K-EGTA, pH 7.3. ATP and ADP were added as the potassium salts. All currents were measured at membrane potential of −50 mV (pipette voltage = 50 mV) at room temperature. Data were analyzed using pCLAMP8 and Microsoft Excel.

Data were presented as means ± SE. Statistical analysis was performed using independent two-population two-tailed Student's t test, with P < 0.05 considered statistically significant.

Previously, we reported several CHI-associated SUR1 mutations that cause endoplasmic reticulum retention and reduced surface expression of K_ATP channels (12,14,16). To extend these studies, we screened 15 additional SUR1 missense mutations that are distributed throughout the protein for their potential effects on surface expression of the channel. The mutations are divided into three groups based on locations according to the current SUR1 topology model (Fig. 1). The first group, including G7R, N24K, F27S, R74W, and E128K, is located in the first transmembrane domain TMD0; the second group, including R495Q, E501K, L503P, F686S, and G716V, is located in the second transmembrane domain TMD1 extending through the first nucleotide binding domain; the third group, including K1337N, L1350Q, S1387F, L1390P, and D1472H, is clustered in the second nucleotide binding domain and the COOH terminus of the protein. More than one-half of these mutations have been previously identified (for references, see Table 1) (22), although detailed analysis of how these mutations affect K_ATP channel biology is still lacking. The mutations that have not been previously reported in the literature include N24K, E128K, and L1390P. Clinical and genetic information available on patients carrying these mutations are summarized in Table 1.

Cell surface expression of K_ATP channels was analyzed using immunoblotting, immunofluorescent staining, chemiluminescence assays, and electrophysiological recording of channel activity; these methods have been documented extensively in previous studies and provide both qualitative and quantitative measures of channel expression at the cell surface (12,14,16,25,26). All of the mutations were engineered into a SUR1 construct tagged at the extracellular NH₂-terminus with a FLAG epitope (fSUR1) to facilitate detection at the cell surface; the FLAG tag has no discernable effect on either expression or function of wild-type channels (12). In immunoblots, wild-type fSUR1 coexpressed with Kir6.2 is detected as a core-glycosylated form of ∼140 kDa and an complex-glycosylated form of ∼180 kDa. Because complex glycosylation occurs in the medial Golgi, the upper fSUR1 band indicates fSUR1 that has passed the endoplasmic reticulum quality control checkpoint by forming channel complex with Kir6.2. The immunoblotting analysis revealed reduced level of the upper band in many SUR1 mutations (only selective mutants are shown in Fig. 2A), suggesting that these mutations might cause channel trafficking defects. To further investigate the subcellular localization of the mutants, we performed immunofluorescent staining of fSUR1 present at the cell surface and the total fSUR1 present in the cell (see research design and methods). The same mutations that exhibited no or reduced mature band in immunoblotting analysis also showed no or reduced surface fluorescence signals; total fluorescent staining revealed that the mutant proteins were present but were retained intracellularly (two mutants are shown in Fig. 2B; see also Supplemental Fig. 1). To better quantify channel expression level at the cell surface, we labeled surface fSUR1 with the anti-FLAG antibody, which was then detected using a chemiluminescence-based enzymatic reaction (14,16). Results from this assay showed that F27S, R74W, E128K, R495Q, E501K, L503P, F686S, G716V, L1350Q, and D1472H mutant channels had greatly reduced surface expression (<20% of wild-type level)—whereas G7R and N24K mutant channels displayed modestly decreased surface expression level (>30% but <50% of wild-type level) and K1337N, S1378F, and L1390P exhibited normal or mildly reduced expression (>60% of wild-type level; Fig. 3A). Channel density of the various mutants estimated using inside-out patch-clamp recording is in general agreement with chemiluminescence assay results (Fig. 3B; Supplemental Fig. 2), providing confirmation that the surface fSUR1 detected was co-assembled with Kir6.2 as K_ATP channel complex. Together, the above analyses identify new mutations of the group that likely cause hyperinsulinism by at least in part reducing surface expression of functional K_ATP channels.

In addition to loss of channel expression at the plasma membrane, inability of the channel to respond to MgADP stimulation is another prominent molecular defect that accounts for lack of K_ATP channel activity in hypoglycemic CHI patients (27,28). We therefore examined how those mutants that have good expression at the cell surface, including K1337N, S1387F, and L1390P, respond to MgADP using inside-out patch-clamp recording (27,28). In wild-type channels, exposure of the cytoplasmic face of the channel to 1 or 0.1 mmol/l ATP caused near total inhibition of channel activity; however, addition of 0.5 mmol/l MgADP to 0.1 mmol/l ATP antagonized the inhibition by 0.1 mmol/l ATP, demonstrating the stimulatory effect of MgADP on channel activity. This response to MgADP was abolished in the S1387F and L1390P mutations and slightly reduced (not statistically significant) in the K1337N mutation. Representative recordings from these mutants are shown in Fig. 4A, and the averaged data are summarized in Fig. 4B.

The observation that mutant SUR1 proteins lacking surface expression are present intracellularly prompted us to test whether the mutant proteins could be rescued to the cell surface using pharmacological agents that have been reported to correct channel trafficking defects (13,16). We first examined the effects of sulfonylureas, which were shown previously to improve surface expression of the A116P- and V187D-SUR1 mutants (16), on the trafficking mutants identified in this study (those that had surface expression <50% of wild type based on chemiluminescence assays shown in Fig. 3A). In Western blots, several mutants, including G7R, N24K, F27S, R74W, and E128K, all located in TMD0, exhibited increased complex-glycosylated SUR1 in cells coexpressing Kir6.2 on overnight treatment with 1 μmol/l glibenclamide (Fig. 5A). To determine whether the increased complex-glycosylated SUR1 corresponds to increased cell surface expression of mutant channels, immunofluorescent staining of surface fSUR1 was performed in cells treated with 1 μmol/l glibenclamide or DMSO control. As shown in Fig. 5B, glibenclamide treatment markedly increased surface expression of the aforementioned TMD0 mutants. Chemiluminescence assays yielded consistent results, with the surface expression level increased by 30–70% (Fig. 5C). Tolbutamide (at 300 μmol/l) similarly improved surface expression of the five TMD0 mutants in all three different assays (data not shown). To ensure that the trafficking defects of TMD0 mutations and their rescue by sulfonylureas are also seen in a cellular environment in which the channels normally reside, we examined whether A116P, a TMD0 trafficking mutant we documented previously, behaves the same in the insulin-secreting cell line INS-1 as in COS cells. Because INS-1 cells are difficult to transfect, we resorted to the adenoviral system for expression of the mutant protein (see online appendix for detailed procedures) (29). As shown in Fig. 5D, A116P-SUR1 expressed in INS-1 cells failed to mature into the complex-glycosylated form; however, overnight treatment of cells with 1 μmol/l glibenclamide overcame this processing defect and led to the appearance of the complex-glycosylated form. These observations recapitulate what we have reported previously for A116P-SUR1 expressed in COS cells (16), suggesting that our findings in COS cells are likely applicable to insulin-secreting cells.

We next examined whether diazoxide, a K⁺ channel opener that also binds SUR1, is able to rescue our trafficking mutants. In contrast to sulfonylureas, diazoxide (100 μmol/l) did not have significant rescue effects on surface expression of any of the trafficking mutants either in immunostaining experiments (not shown) or in chemiluminescence assays (Fig. 6). Because diazoxide has been reported by another group to correct channel trafficking defect caused by the R1394H-SUR1 mutation (13), we examined the expression pattern of the mutant and how it responds to diazoxide. In our hands, the R1394H mutant channels have expression levels very comparable with wild-type channels. The surface expression level is ∼80% that of wild type, and diazoxide had no obvious effect on either expression level or cellular distribution of the mutant (data not shown).

The key defect in CHI caused by K_ATP channel mutations is inadequate channel activation by hypoglycemia. For channels expressed at the cell surface to perform their physiological function, they have to be able to open when cell metabolism is inhibited. We sought to determine whether the TMD0 mutant channels rescued to the plasma membrane are responsive to the metabolic signals associated with hypoglycemia. Channel activity in intact cells was assessed using the ⁸⁶Rb⁺ efflux assay, which is well established for monitoring channel response to metabolic inhibition (23,28,30). In cells expressing wild-type K_ATP channel subunits, >90% efflux was observed in a 40-min time period after cells were preincubated for 30 min with the metabolic inhibitors 2-deoxy-d-glucose (1 mmol/l) and oligomycin (2.5 μg/ml). In contrast, untransfected cells showed <10% efflux in 40 min (Fig. 7). Consistent with the mutations causing hyperinsulinism, cells expressing one of the five TMD0 mutant channels exhibited reduced Rb⁺ efflux compared with cells expressing wild-type channels. To test whether sulfonylurea treatment improves the response of cells expressing the mutants to metabolic inhibition, we treated the cells with 300 μmol/l tolbutamide overnight and removed the tolbutamide 4 h before the assay. Tolbutamide was chosen over glibenclamide for this experiment because we have previously shown that glibenclamide, which binds SUR1 with high affinity, remains bound to channels rescued to the cell surface, preventing the channel from opening on metabolic inhibition; tolbutamide can be easily washed off to recover channel function (16,31). Cells pretreated with tolbutamide followed by subsequent removal showed greater efflux than untreated cells (Fig. 7; see also Supplemental Fig. 1). Our results demonstrate that reversible sulfonylurea drugs, such as tolbutamide, can be used to correct K_ATP channel trafficking defects and recover channel function in CHI patients carrying the TMD0-SUR1 mutations described in our earlier study (16) and the present study.

K_ATP channel mutations that lead to partial or total loss of channel function are a major cause of CHI (4,6,11,32,33). Understanding how mutations affect molecular and cellular properties of the channel to blunt channel function may help develop new disease treatment strategies. Several reports have documented decreased K_ATP channel expression at the plasma membrane as mutation-induced defects that account for loss of channel function (14–16,25,33,34). Importantly, defective surface expression seen in some of these mutant channels may be corrected pharmacologically with the resulting channels being physiologically functional (13,16), raising the possibility of a novel treatment for patients carrying these mutations. In this study, by screening 15 disease-associated SUR1 missense mutations, we identified 12 additional SUR1 mutations that reduce K_ATP channel surface expression by >50%. Of these, five responded to sulfonylurea treatment with significantly improved surface channel expression. Moreover, mutant channels that were rescued by tolbutamide all showed improved activity in response to metabolic inhibition in ⁸⁶Rb⁺ efflux assays, demonstrating restoration of channel function. Our findings reinforce the notion that mislocalization of K_ATP channels is a major mechanism contributing to loss of channel function in CHI and expand the number of mutations that may be targeted by pharmacological chaperones to restore channel expression and function.

The first two trafficking mutations that we reported to be rescued by sulfonylurea drugs are A116P and V187D, both located in TMD0 of SUR1 (16). Strikingly, of the 12 new trafficking mutants that we identified in this study, only the TMD0 mutants responded to sulfonylurea rescue. TMD0 is a distinct structural feature of the SUR protein that separates it from other prototype ATP-binding cassette (ABC) transporter family members, such as CFTR (cystic fibrosis transmembrane conductance regulator) and P-glycoprotein, which contain only two transmembrane domains. TMD0 alone has been shown to associate with Kir6.2 and modulate its gating property, indicating that the domain is in and of itself a structural and functional entity (34,35). Recently, we have shown that sulfonylureas rescue TMD0 trafficking mutants via direct interactions with the channel complex (31). However, rather than binding to TMD0, sulfonylureas appear to bind to regions downstream of TMD0 to exert their chaperoning effects. First, a truncated SUR1 of TMD0 alone containing the A116P or V187D trafficking mutations failed to respond to sulfonylurea rescue. Second, two point mutations, Y230A and S1238Y, that are located downstream of TMD0 and are known to diminish or abolish glibenclamide and tolbutamide binding accordingly affected the ability of the drugs to rescue channel trafficking defects caused by TMD0 mutations (31,36,37). Thus, it appears that sulfonylurea binding to the ABC core region of SUR1 somehow overcomes the defect caused by TMD0 mutations. There is precedent that such substrate-induced transmembrane domain interactions occur: In another ABC transporter, the human P-glycoprotein, substrate-induced interactions between the two transmembrane domains promote superfolding of partially unfolded intermediates into the mature enzyme (38). An additional interesting and relevant finding from our recent study is that the Kir6.2 subunit coexpression is required for sulfonylureas to correct trafficking defects that resulted from SUR1-TMD0 mutations (31). It is possible that the TMD0 trafficking mutations interfere with normal association with Kir6.2 and that sulfonylurea binding promotes SUR1-Kir6.2 interactions to overcome channel biogenesis/trafficking defects.

The mutations that we examined here cause either significant reduction in channel expression level at the cell surface or channel response to MgADP when expressed as homozygous mutants, with the exception of the K1337N mutation, which only exhibited a slight reduction in surface expression and MgADP response that did not reach statistical significance. Therefore, in most cases, the in vitro channel phenotype is consistent with the loss of channel function phenotype in patients. Because the focus of this study is identifying new trafficking mutants and their rescue by pharmacological chaperones, we did not exhaustively analyze all of the mutants for their MgADP response or other channel properties that might affect channel function. However, it is likely that some of the mutations cause both expression and functional defects, as we have documented before (12,14). Also, some mutations are compound heterozygous with another SUR1 mutation in the other allele that may also reduce channel function. For example, both R74W and R495Q are compound heterozygous mutations with R1215Q. Based on in vitro functional phenotypes of mutant hamster SUR1/rat Kir6.2 channels expressed in COS cells, one would expect that in patients, in addition to the expression defects caused by the R74W or R495Q mutation, the R1215Q mutation would also reduce channel function by reducing channel response to MgADP (28). Another example is the combination of D1472H and ΔF1388, both of which result in nearly complete loss of channel expression at the cell surface (12). The functional data also show that in many mutants, channel activity is not completely lost but only partially reduced (Figs. 3, 4, and 7). This is consistent with the brisk acute insulin responses to tolbutamide observed in vivo in some of the diffuse cases of CHI (39). This was particularly the case in a patient with compound heterozygous R74W/R1215Q mutations who had a greater than normal insulin response to tolbutamide. Overall, the in vitro functional analysis provides valuable information by allowing prediction of phenotype based on genotype to facilitate disease diagnosis and management (40).

Despite their defective biogenesis or trafficking, the TMD0 mutants tested so far all retain the ability to respond to metabolic inhibition in the Rb efflux assay, suggesting that they would improve β-cell function once rescued to the cell surface (16). However, it is important to recognize that the experiments described in this study on the identification of trafficking mutants and their response to sulfonylureas were all conducted in COS cells, which differ significantly from β-cells. In particular, recent studies have suggested that K_ATP channels are concentrated in insulin granules, which constitute the regulated secretory pathway absent in COS cells (41,42). Because there are many examples of cell-type–specific protein trafficking regulation (13,21,43), what we found in COS cells may not extrapolate directly to β-cells. In this regard, our results that a previously published TMD0 trafficking mutant, A116P, exhibits the same trafficking defect and response to sulfonylurea rescue in INS-1 cells as in COS cells provide some assurance that the TMD0 mutants are likely to behave similarly in their native environment. Nevertheless, it is imperative to confirm our findings in β-cells for all other trafficking mutants in the future. Another important point to consider is that because sulfonylureas also inhibit K_ATP channel activity, channel function can only be restored after drug removal. Thus, a future challenge in harnessing the sulfonylurea chaperoning effect to benefit CHI patients carrying TMD0 mutations would be to develop sulfonylurea administration strategies or analog compounds that will maximize the chaperoning effect of sulfonylureas but minimize their inhibitory effect on channel activity.

C.A.S. has received National Institutes of Health (NIH) Grants R01-DK-56268 and 5-MO1-RR-000240. S.-L.S. has received NIH Grants R01-DK-57699 and DK-66485.

We thank Jeremy Bushman, Jillene Casey, Malini Kochhar, and Emily Pratt from the Shyng laboratory for technical assistance and comments on the manuscript. We are grateful to the patients, the General Clinical Research Center staff, and the Children's Hospital of Philadelphia nurses for participating in this research.