Glucose-Induced Nuclear Shuttling of ChREBP Is Mediated by Sorcin and Ca²⁺ Ions in Pancreatic β-Cells

Plasmids c-myc–ChREBP(WT) and c-myc–Pdx-1 have been described elsewhere (5,25). Plasmid c-myc–ChREBP(1–242) was generated by digesting c-myc–ChREBP(WT) by XbaI and self-ligation. Plasmids pGBKT7-ChREBP and pChREBP-enhanced green fluorescent protein (EGFP) were generated by subcloning full-length ChREBP cDNA in frame with the Gal4 DNA binding domain and EGFP, respectively, at the Sal1 sites of pGBKT7 and pEGFP-C2 (Clontech). Plasmid sorcin-HA was derived from the yeast vector pGADT7 (see below) using the following primers: forward, 5′-CGA AGC TTA TGG CGT ATC CCG GGC AC-3′ and reverse, 5′-ATG TGC GGC CGC GGC ATA GTC TGG GAC GTC ATA TGG ATA GAC GGT CAT GAC ACA CTG GAT G-3, HA tag underlined, and subcloned at SalI and NotI sites of pcDNA3.1⁺ (Invitrogen). Plasmids −183LPK.luci and −148LPK.luci have been described in Ref. 26. Plasmid −1081TXNIP.luci was obtained from Addgene and described in Ref. 27. Adenovirus sorcin-HA, also expressing GFP from a single bicistronic mRNA under the cytomegalovirus (CMV) promoter, was generated by three-piece ligation of sorcin-HA and IRES-EGFP in pShuttle-CMV (28) using BglII, XhoI, and XbaI restriction sites. Adenovirus production was then performed as in Ref. 28. Sorcin (sc-41017-v) and control (sc-108080) short hairpin RNA (shRNA) mouse lentiviruses were purchased from Santa Cruz. All constructs were verified by DNA sequencing.

Human embryonic kidney 293, MIN6, and INS1 (832/13) cells were cultured as in Ref. 5 unless specified otherwise. Cells were transfected using Lipofectamine 2000 for plasmids and RNAiMAX for small interfering RNA (siRNA) (final concentration 2.5 to 5 nM) in Opti-Mem (Invitrogen).

MIN6 cDNA library was generated according to Matchmaker Library Construction and Screening Kit (Clontech) using Oligo d(T) primers, 2.0 μg total RNA, yeast strain AH109, and pGADT7-Rec AD. The transformants were selected on SD/-leu plates before yeast mating.

Yeast two-hybrid was performed according to Matchmaker Gal4 Two-Hybrid System 3 (Clontech). In brief, yeast mating was performed using yeast strain AH109 containing the MIN6 cDNA library, and yeast strain Y187 was pretransformed with pGBKT7-ChREBP. Screens were conducted at high (-adenine/-histidine/-leucine/-tryptophan) stringency with 5 mmol/L 3-AT to prevent background growth. The library cDNAs contained in pGADT7 were identified by sequencing.

Cells were transfected with plasmid vectors encoding c-myc–ChREBP(WT), c-myc–ChREBP(1–242), c-myc–Pdx-1 and sorcin-HA, or sorcin-HA alone for 48 h before coimmunoprecipitation with mouse anti–c-myc and Western blotting as described in Ref. 29 using either mouse anti–c-myc, mouse anti-HA, or mouse anti-sorcin antibodies.

MIN6 β-cells were cultured for 16 h in 3 mmol/L glucose followed by 3 h in 3 or 30 mmol/L glucose. The cells were washed twice with PBS and incubated with 1 mmol/L disuccinimidyl suberate (Pierce), at room temperature, for 30 min. The crosslinker reaction was quenched by incubating with 20 mmol/L Tris (pH 7.4) for 15 min at room temperature, followed by washing twice with ice-cold PBS. The cells were lysed in Nonidet-40 lysis buffer before immunoprecipitation with anti-sorcin antibody and Western blotting as above.

MIN6 β-cells were transfected with scrambled or specific sorcin siRNA for 48 h. Chromatin immunoprecipitation was performed using anti-ChREBP antibody and mouse L-PK promoter primers as described (5,6).

MIN6 β-cells, nontransfected or transfected with sorcin-HA for 48 h, were cultured for 16 h in 3 mmol/L glucose followed by 6 h in 3 or 30 mmol/L glucose. The cells were scraped in ice-cold PBS, followed by centrifuging at 500g for 2 min. The cell pellets were lysed, and nuclear proteins were separated from the cytosolic fraction using the NucBuster Protein Extraction Kit (Novagen) before Western blotting.

MIN6 β-cells seeded in 24-well plates were transfected with 1.5 μg of either −183LPK.luci, −148LPK.luci, or −1081TXNIP.luci reporter plasmids, 0.5 μg of either GFP or sorcin cDNA, and 1.5 ng of renilla luciferase to correct for transfection efficiency, using Lipofectamine 2000 (Invitrogen) and OPTI-MEM (5 mmol/L glucose) for 24 h followed by 25 mmol/L glucose Dulbecco’s modified Eagle’s medium for 24 h before cell lysis and luciferase assay using Dual-Luciferase Reporter Assay System (Promega).

MIN6 β-cells were fixed in 4% (volume for volume) paraformaldehyde, permeabilized with 0.3% Triton X-100, and stained with rabbit anti–c-myc antibody (1:250; Santa Cruz), and mouse anti-HA antibody (1:1000; 12CA5, Roche), followed by Alexa 488 conjugated goat anti-rabbit IgG and Alexa 568 conjugated donkey anti-mouse IgG (1:1,000; Invitrogen). Cells were imaged using a Zeiss Axoivert 200M microscope fitted with a PlanApo ×63 oil immersion objective. Confocality was achieved using a Nipkow spinning disc under the control of Volocity 4.0 (Improvision) software.

MIN6 β-cells were loaded with fura-2 a.m. (2 μmol/L in DMSO) in Krebs-Ringer bicarbonate HEPES (KRBH) buffer (125 mmol/L NaCl, 3.5 mmol/L KCl, 1.5 mmol/L CaCl₂, 0.5 mmol/L MgSO₄, 0.5 mmol/L NaH₂PO₄, 25 mmol/L NaHCO₃, and 10 mmol/L HEPES), supplemented with 3 mmol/L glucose for 60 min at 37°C. Cells were perifused in KRBH and imaged at 37°C using an Olympus IX-70 with an Imago charge-coupled device camera (Till Photonics, Grafelfing, Germany) controlled by TILLvisION software (Till Photonics). Cells were illuminated alternately for 5 ms at 340 and 380 nm, and the emission was filtered at 510 nm. After background and auto-fluorescence were subtracted (by quenching with 10 mmol/L MnCl₂ and 1 μmol/L ionomycin), free cystolic intracellular Ca²⁺ ([Ca²⁺]_cyt) was calculated using a look-up table created from standard Ca²⁺ solutions (Molecular Probes).

MIN6 β-cells were transfected either with ChREBP-EGFP plasmid alone or with sorcin siRNA for 48 h. Cells were then incubated overnight with 3 mmol/L glucose Dulbecco’s modified Eagle’s medium, and subsequently the coverslips were rinsed with 3 mmol/L glucose KRBH buffer and mounted on a hot stage with continuous perifusion with warm KRBH. Cells were imaged using a Zeiss Axiovert 200M microscope (Carl Zeiss, Jena, Germany) fitted with a PlanApo 63× oil-immersion objective and a 1.5 Optivar attached to a Nokigawa spinning disc confocal head. Hoechst (Invitrogen) staining of nuclei was visualized through DAPI filter. GFP-expressing samples were illuminated using steady-state 488 nm laser line, and emission was collected through ET535/30 emission filter (Chroma) for every 5 s. Images were captured using a Hamamatsu EM charge-coupled device digital camera, model C9100-13, controlled by Volocity 4.0 software. Deconvoluted images were analyzed by obtaining the gray values for every time point in the region of interest drawn around nucleus and the cytosol, which was then corrected for background and normalized, and ratios of nuclear to cytosolic distribution were generated.

Mouse islets were isolated (5) and partially dissociated by incubation in Hank’s based cell dissociation buffer (Invitrogen) for 10 min by gentle pipetting. To assess the effects of intracellular Ca²⁺, primary cell clusters were transduced with c-myc–ChREBP-expressing adenoviral particles (also expressing GFP) (5) at a multiplicity of infection of ∼50 for 6 h, followed by 48 h culture in fresh medium. Dissociated islets were then incubated overnight in 3 mmol/L glucose and thereafter in 3 or 17 mmol/L glucose in KRBH buffer containing either 1.5 mmol/L CaCl₂ or 10 mmol/L EGTA for 30 min and fixed in 4% paraformaldehyde for 10 min. After permeabilization with 0.1% Triton × 100, immunostaining was carried out using the following antibodies: monoclonal anti–c-myc antibody (1:200; Roche), polyclonal anti-swine insulin (1:200; DakoCytomation, Ely, UK), and Alexa 564 or 633-coupled secondary antibodies (1:1,000; Molecular Probes), and coverslips were mounted in VECTASHIELD (Vector Laboratories) hardset medium with DAPI. Imaging was performed using a Leica SP2 upright confocal microscope with an HCX PL APO 63× 1.32 numerical aperture oil immersion objective and the following laser lines: Argon ion (488 nm), HeNe (633 nm), 561 nm diode, and ultraviolet (350 nm) controlled by Leica LAS AF Lite software. Acquired images were analyzed using Volocity software (Improvision). To assess the effect of sorcin overexpression or silencing on endogenous ChREBP nuclear localization, dissociated islets were transduced with either sorcin (also expressing GFP) or GFP adenoviral particles as above or with shRNA sorcin or control lentiviral particles (Santa Cruz) according to the manufacturer’s instructions in 11 mmol/L glucose RPMI 1640 for 48 h followed by 1 h in 3 mmol/L glucose RPMI 1640 and 6 h in 17 mmol/L glucose RPMI 1640. Cells were then fixed and immunostained using rabbit anti-ChREBP (1:100; Thermo Scientific) and mouse anti-sorcin (1:50; Invitrogen) and secondary goat–anti-rabbit Alexa 568, and goat–anti-mouse Alexa 633 (both at 1:1,000; Invitrogen).

MIN6 β-cells were transfected with sorcin-specific or scrambled siRNAs for 48 h before insulin secretion assay (5).

Data are given as mean ± SEM of at least three individual experiments. Comparisons between means were performed using two-tailed Student t test for unpaired data, with Bonferroni correction and ANOVA.

We first detected sorcin as a ChREBP interactor in a yeast two-hybrid screen using full length ChREBP as bait in a pancreatic β-cell MIN6 library (not shown). We confirmed the interaction in mammalian cells by coimmunoprecipitation experiments of overexpressed c-myc–tagged ChREBP and HA-tagged sorcin in human embryonic kidney 293 cells (Supplementary Fig. 1) and INS-1(832/13) β-cells (Fig. 1A, first lane from left). This interaction was further delimited to the first 242 amino acids of ChREBP (Fig. 1A, second lane), a region known as the low-glucose inhibitory domain (12). Establishing the specificity of the ChREBP-sorcin interaction, coimmunoprecipitation experiments did not reveal any interaction between sorcin-HA and c-myc–Pdx-1 (Fig. 1A, fourth lane), a transcription factor that also translocates to the nucleus in response to high glucose in this cell type (25).

We next explored the interaction between endogenous ChREBP and sorcin in MIN6 cells. Such interaction was only detectable after crosslinking experiments, as shown in Fig. 1B. Indeed, crosslinking and immunoprecipitation with anti-sorcin antibody revealed a high molecular complex of ∼140 kDa using both anti-sorcin (Fig. 1B, top panel, second and third lanes) and anti-ChREBP (Fig. 1B, bottom panel, second and third lanes) antibodies. We believe this complex to represent sorcin dimers and ChREBP, since sorcin dimers were also detected after addition of the crosslinker (Fig. 1B, middle panel, second and third lanes).

It has been shown previously that ChREBP constantly shuttles between the cytosol and the nucleus (13). The cytosolic retention domain has been mapped to the NH₂-terminal region of ChREBP, since NH₂-terminal deletion mutants lacking the first 252 amino acids are predominantly nuclear (12,30,31).

Because sorcin interacts with ChREBP specifically through this domain (see Fig. 1A), we next investigated whether sorcin may play a role in controlling the subcellular localization of ChREBP. With the use of spinning disc confocal microscopy, c-myc–ChREBP immunoreactivity was examined in MIN6 β-cells transfected with either c-myc–ChREBP alone or cotransfected with c-myc–ChREBP and sorcin-HA and incubated at low or high glucose concentrations. As expected from the coimmunoprecipitation experiments presented above, ChREBP and sorcin immunoreactivities extensively colocalized, especially at low glucose concentrations (Fig. 2B). Strikingly, however, sorcin overexpression greatly diminished the abundance of ChREBP staining in the nucleus from 25.7 ± 2.1% in the control cells to 9.4 ± 0.8% in the sorcin overexpressed cells (P < 0.005, Fig. 2C), suggesting that sorcin may serve as a cytosolic anchor for ChREBP. This was also observed in cell fractionation experiments where sorcin overexpression reduced ChREBP immunoreactivity in the nuclear compartment of cells incubated in elevated glucose concentration (Fig. 2D). Moreover, sorcin overexpression significantly reduced the activity of the −183L-PK and TXNIP promoters, two well-characterized ChREBP target genes (6,9) in luciferase reporter assays, whereas a truncated L-PK promoter (−148L-PK), lacking the ChREBP binding site, was unaffected by sorcin (Fig. 2E).

We next examined the effects of sorcin silencing on ChREBP binding to the L-PK promoter by chromatin immunoprecipitation (ChIP) and subcellular localization in response to glucose. Using siRNA, we achieved a reduction of ∼86% of sorcin immunoreactivity (Fig. 3A and B). As reported previously (6), glucose (30 vs. 3 mmol/L) stimulation increased ChREBP binding on the proximal L-PK promoter (Fig. 3C, top panel). In contrast, in sorcin-silenced cells, ChREBP binding to the L-PK promoter was constitutive, irrespective of glucose concentrations (Fig. 3C, bottom panel, and Fig. 3D for quantification). Accordingly, inactivation of sorcin by RNAi led to an increase in the number of cells displaying nuclear ChREBP-EGFP fluorescence at resting (3 mmol/L) glucose concentration (83.3 ± 2.9% vs. 23.3 ± 4.0%, P < 0.001; Fig. 3E and F).

These results designate sorcin as a physiological regulator of ChREBP activity in response to glucose in MIN6 cells.

Because sorcin undergoes conformational changes after Ca²⁺ binding (18), we next tested whether Ca²⁺ influx in response to glucose or other agents would drive ChREBP into the nucleus. Imaged dynamically in living MIN6 β-cells, elevated glucose concentrations caused a rapid translocation of ChREBP-EGFP from the cytosol to the nucleus (Fig. 4A and G and Supplementary Movie 1). Parallel single-cell measurements of [Ca²⁺]_cyt revealed that nuclear accumulation of ChREBP closely followed (∼1 min) a glucose-evoked Ca²⁺ signal (Δ[Ca²⁺]_cyt = 312 ± 14 nmol/L; Fig. 3G). Chelation of extracellular Ca²⁺ with 1 mmol/L EGTA (Fig. 4B and G and Supplementary Movie 2) abolished glucose-induced Ca²⁺ entry and thereby Ca²⁺-induced nuclear translocation of ChREBP, which was also suppressed by 10 μmol/L nifedipine, an L-type Ca²⁺ channel blocker, and by 250 μmol/L diazoxide, an activator of ATP-sensitive K⁺ channels (Fig. 4C and G and Supplementary Movie 3) that prevent membrane depolarization and subsequent Ca²⁺ entry. Furthermore, a transient Ca²⁺ peak (Δ[Ca²⁺]_cyt = 108 ± 11 nmol/L) elicited by 250 μmol/L extracellular ATP, an activator of G protein–coupled P2 purinergic receptors (32), led to robust cytosol-nuclear shuttling of ChREBP in the absence of extracellular Ca²⁺ (Fig. 4D and H and Supplementary Movie 4), suggesting that mobilization of intracellular Ca²⁺ stores was sufficient for translocation. Apparently, nuclear translocation of ChREBP also depended upon the presence of elevated concentrations (30 mmol/L) of glucose (Fig. 4E and Supplementary Movie 5), but this largely reflected the increased Ca²⁺ release (Δ[Ca²⁺]_cyt = 31 ± 7 nmol/L at low vs. 108 ± 11 nmol/L at high glucose) in these conditions (Fig. 4H).

We next examined the effects of sorcin silencing on ChREBP nuclear translocation in response to elevated glucose concentrations and Ca²⁺ signals. Although glucose-induced Δ[Ca²⁺]_cyt were not significantly different (330 ± 20 vs. 280 ± 22 nmol/L for control and sorcin-silenced cells, P = 0.55), glucose stimulation failed to trigger nuclear shuttling of ChREBP in sorcin-silenced cells (Fig. 4F and Supplementary Movie 6). On the contrary, a high level of ChREBP was already detected in the nucleus, even before treatments, pointing to the role of sorcin as a ChREBP cytosolic anchor as discussed above. Silencing of sorcin also completely abolished ATP-evoked Ca²⁺ release from intracellular stores (Δ[Ca²⁺]_cyt = 234 ± 20 nmol/L for control vs. −54 ± 18 nmol/L for sorcin-siRNA; Fig. 4I).

We next examined the impact of sorcin inhibition on GSIS. As shown in Fig. 5, sorcin silencing significantly decreased GSIS in pancreatic MIN6 β-cells.

We next examined the effects of Ca²⁺ influx and sorcin in primary β-cells. Dissociated CD1 mouse islets were transduced with c-myc–tagged ChREBP-expressing adenoviral particles (5) for 48 h and incubated overnight in 3 mmol/L glucose and thereafter in 3 or 17 mmol/L glucose KRBH containing either 1.5 mmol/L CaCl₂ or 10 mmol/L EGTA for 30 min before immunostaining and imaging as described in 2research design and methods. The identity of the β-cells was established by simultaneous immunostaining with insulin antibody. As shown in Fig. 6, inhibiting Ca²⁺ influx also abolished ChREBP nuclear localization in response to glucose.

Extending the physiological relevance of the above findings, sorcin overexpression in dissociated primary mouse islets (comprising 70–80% β-cells) (33) diminished endogenous ChREBP nuclear immunoreactivity from 28.6 ± 1.5% in control cells to 16.5 ± 1.0% in sorcin overexpressing cells incubated in elevated (17 mmol/L) glucose (P < 0.001, Fig. 7A and B). Conversely, sorcin silencing increased endogenous ChREBP nuclear staining to 25.5 ± 1.4% compared with 13.6 ± 1.1% in control cells (P < 0.001, Fig. 7C and D), incubated in low (3 mmol/L) glucose.

The chief goal of the current study was to elucidate the molecular mechanisms through which elevated glucose concentrations lead to the nuclear accumulation of ChREBP in the pancreatic β-cell. We show that sorcin provides a Ca²⁺-regulated cytosolic anchor, likely to play an important role in the regulation of this process in both clonal and primary β-cells.

In the liver, glucose-dependent ChREBP nuclear translocation is subject to a series of phosphorylation/dephosphorylation reactions. In fasting and in low glucose conditions, ChREBP is phosphorylated by PKA on S196 and T666 and by AMP-activated protein kinase on S568, which promote cytosol retention and inhibit DNA binding. By contrast, in high glucose conditions, protein phosphatase 2A activation by xylulose 5-phosphate (34) induces dephosphorylation of S196, T666, and S568; nuclear import; and DNA binding (35,36). Dephosphorylation of these residues, however, was not sufficient for ChREBP nuclear translocation in the glucose-responsive pancreatic β-cell line INS1 (832/13) (13). This is unsurprising, since, in the pancreatic β-cell, elevated glucose concentrations are rather associated with protein phosphatase 2A inhibition (37,38) and increased intracellular cAMP levels and PKA signaling (39,40), pointing to the existence of additional mechanisms, possibly including the inhibition by glucose of AMP-activated protein kinase (41).

Here we show that, in the excitable pancreatic β-cell, elevated glucose concentrations provoke ChREBP nuclear translocation through stimulation of Ca²⁺ influx. Thus, extracellular Ca²⁺ chelation with EGTA or pharmacological blockade of Ca²⁺ entry with diazoxide or nifedipine completely abolished glucose-induced ChREBP nuclear shuttling, both in clonal MIN6 (Fig. 4) and primary mouse β-cells (Fig. 6). Conversely, mobilization of intracellular Ca²⁺ with extracellular ATP was sufficient to cause ChREBP nuclear translocation. Together, these data point to a central role for [Ca²⁺]_cyt in this phenomenon. Critically, we show that overexpression or silencing of sorcin in clonal (Fig. 2–4) or primary (Fig. 7) β-cells blocks and inhibits, respectively, nuclear accumulation of ChREBP, implicating sorcin as a likely sensor of cytosolic Ca²⁺.

Changes in [Ca²⁺]_cyt have been reported previously to be detected and decoded as nuclear events via mechanisms involving cytosol-nuclear translocation of nuclear factor activator of transcription and nuclear factor-κB cells (NF-κB) (42). The signaling module reported here may represent a similar system, which allows highly localized changes in cytosolic Ca²⁺, perhaps close to the ER or even a specific Ca²⁺ release channel, to be decoded as a transcriptional response. We thus propose a model (Fig. 8) in which sorcin, a penta-EF-hand protein that undergoes Ca²⁺-dependent conformational changes (15–18), binds to and sequesters ChREBP in the cytosol at low glucose, through an interaction with the NH₂-terminal domain of ChREBP. After glucose stimulation and Ca²⁺-induced Ca²⁺ release, sorcin liberates ChREBP, which is then free to translocate to the nucleus. The regulation of ARNT/HIF-1β gene in pancreatic β-cells supports this model of ChREBP activation. Indeed, we have recently shown that ChREBP is a repressor of ARNT/HIF-1β promoter in pancreatic β-cells (5), whereas another group (43) has shown that blocking Ca²⁺ flux from intracellular Ca²⁺ channels induces the expression of ARNT/HIF-1β in this cell type.

Other interesting findings of this study are the apparent depletion of intracellular Ca²⁺ stores (Fig. 4I) and the inhibition of GSIS (Fig. 5) after sorcin silencing in MIN6 cells. The absence of sorcin may lead to a depletion of intracellular Ca²⁺ stores by allowing random activation of RyR (21). Another possibility would be a change in gene expression, possibly the expression of the G protein–coupled P2 purinergic receptors, hence diminishing the response to extracellular ATP. However, any such changes were apparently efficiently countered by modifications to Ca²⁺ pumping since neither resting nor glucose-stimulated increases in cytosolic Ca²⁺ were perturbed by the knockdown of sorcin (Fig. 4I). It follows that the interaction of sorcin with voltage-dependent (L-type) Ca²⁺ channels (20) in the β-cell may be of limited importance in the control of Ca²⁺ influx. The latter observation argues against substantial changes in β-cell glucose metabolism or in cellular viability after sorcin silencing in the short period (48 h) of our study. However, a chronic sorcin inhibition is likely to induce changes in gene expression through a constitutive ChREBP signaling pathway, including downregulation of ARNT/HIF-1β (5) and upregulation of TxNIP (9). Although the latter is often associated with oxidative stress and cell death by mechanisms involving mitochondrial Ca²⁺ accumulation (44), we would note that the possible depletion of ER Ca²⁺ stores (see above), required in several cell types for efficient accumulation of mitochondrial Ca²⁺ (45), may reduce the magnitude of any proapoptotic effects of TxNIP induction. Further detailed studies will be required in the future to discriminate between these possibilities.

An event distal to increase in [Ca²⁺]_cyt is therefore likely to be involved in the inhibition of GSIS. This may, for example, include the less efficient mobilization or fusion of secretory granules with the plasma membrane in sorcin-silenced cells (46).

An exciting corollary of the work presented here is the development of novel therapeutic approaches involving sorcin activators or sorcin overexpression in the treatment of failing β-cells in type 2 diabetes, which may complement the positive action of such agents in the heart (47,48).

This work was supported by grants from Diabetes UK (BDA:RD04/0002895) and the Wellcome Trust (WT082366MA) to I.L. and by the Wellcome Trust (Programme Grant 081958/2/07/Z), the European Community's Seventh Framework Programme (FP7/2007-2013) for the Innovative Medicine Initiative under grant agreement n° 11500 (IMIDIA) to G.A.R.

No potential conflicts of interest relevant to this article were reported.

N.A.N., G.M., and I.L. designed the experiments, researched data, wrote the manuscript, and contributed to the discussion. G.A.R. designed the experiments, reviewed and edited the manuscript, and contributed to the discussion. As the corresponding author and guarantor of this article, I.L. takes full responsibility for the work as a whole, including the study design, access to data, and the decision to submit and publish the manuscript.

This study was presented as an oral communication at the 46th Annual Meeting of the European Association for the Study of Diabetes, Stockholm, Sweden, 20–24 September 2010; the Congrès de la Société Francophone de Diabète, Geneva, Switzerland, 22–25 March 2011; the Annual Professional Conference of Diabetes UK, London, U.K., 30 March–1 April 2011; and the 28e Congrès de la Société Française d’Endocrinologie, Clermont-Ferrand, France, 10–13 October 2011.

The authors thank Prof. Jean-Claude Henquin, University of Louvain Faculty of Medicine, Brussels, Belgium, for helpful discussion.