Inhibition of Pancreatic β-Cell Ca2+/Calmodulin-dependent Protein Kinase II Reduces Glucose-stimulated Calcium Influx and Insulin Secretion, Impairing Glucose Tolerance*

Background: Glucose activates CaMKII in β-cells, how this influences glucose homeostasis has not been determined. Results: Inhibiting CaMKII in mouse β-cells causes glucose intolerance by reducing Ca2+ entry and insulin secretion. Conclusion: CaMKII is a β-cell Ca2+ sensor that amplifies secretagogue-induced Ca2+ entry and insulin secretion to maintain glucose homeostasis. Significance: This provides the first evidence that β-cell CaMKII modulates glucose homeostasis under physiological and insulin resistant states. Glucose-stimulated insulin secretion (GSIS) from pancreatic β-cells is caused by Ca2+ entry via voltage-dependent Ca2+ channels. CaMKII is a key mediator and feedback regulator of Ca2+ signaling in many tissues, but its role in β-cells is poorly understood, especially in vivo. Here, we report that mice with conditional inhibition of CaMKII in β-cells show significantly impaired glucose tolerance due to decreased GSIS. Moreover, β-cell CaMKII inhibition dramatically exacerbates glucose intolerance following exposure to a high fat diet. The impairment of islet GSIS by β-cell CaMKII inhibition is not accompanied by changes in either glucose metabolism or the activities of KATP and voltage-gated potassium channels. However, glucose-stimulated Ca2+ entry via voltage-dependent Ca2+ channels is reduced in islet β-cells with CaMKII inhibition, as well as in primary wild-type β-cells treated with a peptide inhibitor of CaMKII. The levels of basal β-cell cytoplasmic Ca2+ and of endoplasmic reticulum Ca2+ stores are also decreased by CaMKII inhibition. In addition, CaMKII inhibition suppresses glucose-stimulated action potential firing frequency. These results reveal that CaMKII is a Ca2+ sensor with a key role as a feed-forward stimulator of β-cell Ca2+ signals that enhance GSIS under physiological and pathological conditions.

days as indicated. Controls were RIP-rtTA animals or RIP-rtTA ϩ tetO-GFP mice (Jackson ImmunoResearch Laboratories, 018913); as indicated, both were treated with doxycline identically to the EAC3I-GFP mice. For acute inhibition of islet-cell CAMKII, C57BL/6 islets were dispersed into 20-cell clusters and then treated with the cell-permeable CAMKII inhibitor peptide, autocamtide-2 related inhibitory peptide II (AIP2), which is an Antennapedia transport peptide fused to AIP2 (EMD Millipore). Islet cells were incubated with 20 M AIP2 in 2 mM glucose for 20 min immediately before Ca 2ϩ imaging.
Mouse Diets and Glucose Tolerance Testing-Mice were placed either on a normal chow diet or a high fat diet (HFD, 60 kcal% fat; Research Diets, Inc.) and monitored for glucose tolerance. The glucose tolerance test (GTT) was performed as described previously by injecting 2 mg/kg dextrose (animals on a normal chow diet) or 1 mg/kg (animals on a HFD) and monitoring blood glucose at the indicated time points post glucose injection (11). The mice fed a normal chow diet were treated with doxycycline at 6 weeks of age, and GTT was performed at 7.5 weeks of age. A cohort of mice was also placed on a HFD at 3 weeks of age for 2 months when GTT was performed; at this time, doxycycline was added to the drinking water for 14 days in the presence of the HFD following which another GTT was performed.
Mouse Islet and ␤-Cell Isolation-Islets were isolated from pancreata of mice, using collagenase digestion and Ficoll gradients as described previously (12). Islets were plated or dissociated in 0.005% trypsin, placed on glass coverslips, and cultured for 16 h in RPMI 1640 medium supplemented with 10% fetal calf serum, concentrations of glucose-specified, 100 international units ml Ϫ1 penicillin, and 100 mg ml Ϫ1 streptomycin. Dissociated ␤-cells were specifically used in all voltage clamp experiments recording Ca 2ϩ currents. ␤-Cells on the periphery of intact islets were recorded in current clamp mode in all of the membrane potential recordings. Cells and islets were maintained in a humidified incubator at 37°C under an atmosphere of 95% air and 5% CO 2 .
Western Blot Analysis-Mouse islets in groups of 50 were treated with 1 M ionomycin for 2 min. Protein extracts were prepared from islets by extraction with SDS loading buffer (1% SDS, 30 mmol/liter Tris-HCl (pH 6.8), 5% ␤-mercaptoethanol, 5% glycerol, and 0.1% bromphenol blue) with protease and phosphatase inhibitors at 80°C for 10 min. After electrophoresis through a 4 -12% denaturing polyacrylamide gel, proteins were prepared as a Western blot on a nitrocellulose membrane (Bio-Rad). Anti-phosphosynapsin antibody (Santa Cruz Biotechnology) or anti-GAPDH (Rockland Immunochemicals) was used to probe the membrane at 1:250 or 1:700 dilution, respectively, in PBS, 0.1% Tween 20, and 3% powdered dried milk followed by an HRP-coupled secondary antibody (Jackson ImmunoResearch Laboratories) at 1:5000 in the same solution. The membranes were washed in PBS containing 0.1% Tween between and after antibody incubations; HRP was illuminated using Pico Signal (Pierce) and exposed on Kodak X-Omat Blue film.
Perforated Patch Electrophysiology-Patch electrodes (2-4 micro-ohms) loaded with solution containing (in mmol⅐liter Ϫ1 ) 140 KCl, 1 MgCl 2 [H 2 O] 6 , 10 EGTA, 10 HEPES (pH 7.25 with KOH) and the pore-forming antibiotic amphotericin B (Sigma) were used to record islet-attached ␤-cells (13). Islets were perfused with Krebs-Ringer-HEPES buffer containing 119 mmol⅐ liter Ϫ1 NaCl, 2 mmol⅐liter Ϫ1 CaCl 2 , 4.7 mmol⅐liter Ϫ1 KCl, 10 mmol⅐liter Ϫ1 HEPES, 1.2 mmol⅐liter Ϫ1 MgSO 4 , 1.2 mmol⅐liter Ϫ1 KH 2 PO 4 , adjusted to pH 7.35 with NaOH, with the indicated concentrations of glucose and compounds. Cells on the periphery of islets were sealed in voltage clamp at Ϫ80 mV, and access was obtained over several minutes through perforations by amphotericin (13). After being switched to current clamp, cells that had a resting membrane voltage near Ϫ65 mV in 2 mM glucose and fired voltage-dependent Ca 2ϩ channel (VDCC)dependent action potentials post glucose-stimulation were assumed to be ␤-cells.
Whole Measurement of Cytoplasmic Calcium-Islets were incubated (20 min at 37°C) in KRB supplemented with 2 M Fura-2 acetoxymethyl ester (Molecular Probes, Eugene, OR). Fluorescence imaging was performed using a Nikon Eclipse TE2000-U microscope equipped with an epifluorescence illuminator (SUTTER, Inc.) a CCD camera (HQ2, Photometrics, Inc.) and Nikon Elements software (NIKON, Inc.). Cells were perifused at 37°C at a flow of 2 ml/min with appropriate KRB-based solutions that contained glucose concentrations and compounds specified in the figures. Relative Ca 2ϩ concentrations were quantified every 5 s by determining the ratio of emitted fluorescence intensities at excitation wavelengths of 340 and 380 nm (F 340 /F 380 ). The relative Ca 2ϩ is plotted as the average FURA-2 ratio (F 340 /F 380 ) of each experimental group Ϯ the S.E. The averaged area under the curve (AUC) measurement for FURA-2 ratios during the indicated time period (in minutes) was plotted as a bar graph. Data were analyzed using Excel and GraphPad Prism software and compared by Student's t test.
Insulin Secretion Measurements-Mouse islets were allowed to recover following isolation overnight. For insulin measurements, islets were incubated with DMEM with 5.6 mM glucose overnight followed by treatment with the indicated glucose concentrations and analyzed for insulin content using a radioimmunoassay or ELISA-based detection kit (ALPCO Diagnostics, Salem, NH), and data are presented as means Ϯ S.E. For static insulin secretion, islets were incubated in 2 mM glucose containing KRB for 1 h followed by the indicated treatments, and insulin was assayed by ELISA-based detection.
Islet Immunofluorescence Staining-Pancreas cubes from RIP-rtTAϩEAC3I-GFP and RIP-rtTA animals were fixed in 2% paraformaldehyde, paraffin embedded and cut into 5-m sections on a microtome. Islet ␤-cells were stained using insulin (Millipore) and glucagon (Sigma) antibodies at 1:300 in combination with fluorescein isothiocyanate-conjugated secondary antibodies (Jackson ImmunoResearch Laboratories) together with DAPI nuclear stain (Invitrogen).
Islet NAD(P)H fluorescence Imaging-NAD(P)H fluorescence imaging was performed on a LSM780 confocal microscope (Zeiss, Inc.), using a tunable Chameleon Ti:Sapphire laser (Coherent, Inc.). Multiphoton excitation was obtained using a 710-nm wavelength. The emitted light was spectrally resolved using the -mode configuration of the LSM780, with a bandpass of 412-631 nm and a spectral resolution of 8.6 nm. The pixel dwell time was set at 50.4 s, and the pixel size was 0.277 m. We used a Fluor 40ϫ oil objective, with a numerical aperture of 1.30 and dishes with a 0.16 -0.19-mm-thick glass bottom (Mattek). The NAD(P)H emission and GFP emission were extracted by linear unmixing, using ZEN software (Zeiss, Inc.).
The reference spectrum for NAD(P)H was obtained measuring the emission from a freshly made solution of ␤-NADH (Sigma, N8160) in 0.01 M Tris, pH 8.5. The reference spectrum for GFP was measured from a solution of purified GFP in PBS, pH 7.4. The unmixed images were processed using Fiji software (14) to measure the average NAD(P)H intensity across the entire islet and only in the GFP-expressing cells. During the imaging, the islets where maintained in KRBH (Krebs-Ringer bicarbonate/ HEPES buffer), pH 7.40. The islets were kept at 37°C, in a humidified atmosphere with 5% CO 2 , using the microscope stage incubator. The images were collected 8 min post glucose stimulus and 30 m deep in the islet.

CaMKII Activity Modulates Glucose Tolerance by Enhancing
Glucose-stimulated Insulin Secretion-To investigate the physiological role of ␤-cell CaMKII during GSIS, we created a mouse model which allows tetracycline-inducible inhibition of CaMKII. These animals are a cross between mice with a reverse tetracycline transactivator expressed in ␤-cells via an insulin promoter (RIP-rtTA) and mice with a CaMKII pseudosubstrate inhibitor peptide fused to GFP under the control of a tetracycline operator (tetO)-controlled promoter (EAC3I-GFP; Fig.  1A) (10). Doxycycline treatment of these animals, for 1.5 weeks, induces expression of EAC3I-GFP specifically in pancreatic ␤-cells, which can be visualized with GFP fluorescence (Fig.  1B). EAC3I-GFP expression occurs in a majority of ␤-cells, resembling previously published reports on RIP-rtTA mice showing doxycycline induced expression of tetO-controlled transgenes in 68 to 80% of ␤-cells (Fig. 1, B and C) (15,16). white bar is equivalent to 20 M. C, differential interference contrast (DIC) image of an islet cell cluster from an EAC3I-GFP ϩ RIPrtTA mouse on doxycycline for 1.5 weeks stained for GFP (red) and insulin (green), white bar is equivalent to 10 M. D, islet phosphosynapsin 1 (top panels) and GAPDH (lower panels) from islets expressing rtTA or rtTA ϩ EAC3I-GFP in ␤-cells. E, total pancreatic insulin from RIPrtTA or EAC3I-GFP ϩ RIPrtTA mice treated with doxycycline for 1.5 weeks (n ϭ 4 pancreata/group; NS, nonsignificant). F, total islet insulin from RIPrtTA or EAC3I-GFP ϩ RIPrtTA mice treated with doxycycline for 1.5 weeks (n ϭ 4 independent islet preparations/group; NS, nonsignificant). G and H, doxycycline-treated RIPrtTA EAC3I-GFP (G) and doxycycline-treated RIPrtTA (H) pancreatic sections stained for insulin (red), glucagon (green), and nuclei (blue); white bars are equivalent to 20 M. ␤-Cell expression of EAC3I-GFP causes a significant reduction in Ca 2ϩ -induced serine 553 phosphorylation of islet synapsin 1 compared with controls (RIP-rtTA islets), consistent with the effective inhibition of CaMKII activity by EAC3I-GFP expression ( Fig. 1D) (17). However, total pancreatic insulin levels were equivalent between mice with ␤-cell EAC3I-GFP expression and controls (RIP-rtTA pancreata; Fig. 1E). Furthermore, ␤-cell EAC3I-GFP expression does not cause any changes in islet morphology, as ␤-cell size and insulin immunofluorescence were equivalent to controls (Fig. 1, F and G). Therefore, conditional ␤-cell expression of EAC3I-GFP is a good model for assessing the role of CaMKII during physiological insulin secretion.
The effect of ␤-cell CaMKII inhibition on glucose homeostasis was assessed using intraperitoneal (intraperitoneal) glucose tolerance tests. EAC3I-GFP expression in ␤-cells resulted in a significant impairment of glucose tolerance compared with controls (RIP-rtTA mice; Fig. 2A, n ϭ 13 per group). Insulin secretion assays comparing pancreatic islets isolated from control and EAC3I-GFP mice showed that inhibition of ␤-cell CaMKII suppresses insulin secretion (Fig. 2, B-D). Although there was no effect of CaMKII inhibition on basal insulin secretion with 5.6 mM glucose, which resembles blood glucose concentrations during fasting (Fig. 2, B and C), both first and second phase insulin secretion in response to 16.7 mM glucose were reduced in islets with ␤-cell CaMKII inhibition compared with controls (RIP-rtTA islets). Inhibiting ␤-cell CaMKII only significantly reduced GSIS at glucose concentrations Ͼ7 mM, which resemble blood glucose levels (10 mM) reached postprandially (Fig. 2D). In combination, these data show that CaMKII plays a key role in modulating glucose homeostasis and GSIS in vivo.
CaMKII Inhibition Does Not Affect Glucose Metabolism or ␤-Cell K ϩ Channel Activity-GSIS involves multiple steps that may be affected by CaMKII inhibition. First, we examined the potential effect of CaMKII inhibition on ␤-cell glucose metabolism by monitoring NAD(P)H fluorescence under varying glucose conditions (Fig. 3). No significant difference in glucosestimulated changes in NAD(P)H fluorescence was observed between intact islet ␤-cells or whole islets that were obtained from mice with ␤-cell expression of EAC3I-GFP or GFP alone (Fig. 3, A and B). Thus, ␤-cell glucose metabolism is not affected by chronic CaMKII inhibition.
We next investigated potential effects of CaMKII inhibition on the two major ␤-cell K ϩ channels, the delayed rectifier voltage-gated K ϩ channel (K V ) and ATP-sensitive K ϩ channel (K ATP ). The voltage step induction of K V currents was equivalent in control (RIP-rtTA, Fig. 4A) and CaMKII inhibited (EAC3I-GFP, Fig. 4B) ␤-cells, with similar overall current-voltage relationships (Fig. 4C). Similarly, reducing intracellular ATP concentrations induced equivalent K ATP currents in control (RIP-rtTA) and EAC3I-GFP expressing ␤-cells in response  to a voltage ramp (from Ϫ120 to 60 mV, Fig. 4D). Thus, CaMKII inhibition has no detectable affect on K ϩ channel activities in ␤-cells.
Consistent with these findings, insulin secretion induced by 20 mM potassium chloride in the presence of basal glucose, which bypasses ␤-cell glucose metabolism and K ϩ channels by directly depolarizing ␤-cells, was significantly lower from EAC3I-GFP expressing islets compared with controls (Fig. 2, B  and C). Taken together, these data indicate that CaMKII modulates GSIS downstream of glucose metabolism and K ϩ channels to modulate GSIS.
CaMKII Modulates Glucose-stimulated Ca 2ϩ Entry in ␤-Cells-Although CaMKII inhibition reduces Ca 2ϩ -dependent insulin secretion induced by high glucose or direct depolarization (see above), there was no difference in insulin secretion from EAC3I-GFP expressing and control islets in response to the Ca 2ϩ ionophore ionomycin (Fig. 4E) (18), suggesting that CaMKII modulates ␤-cell Ca 2ϩ homeostasis at some level. Consistent with this hypothesis, islet ␤-cells expressing EAC3I-GFP showed reduced basal Ca 2ϩ levels compared with control islets (Fig. 5). Moreover, the changes in Ca 2ϩ following glucose stimulation were also affected: CaMKII inhibition shortened the delay in initiation of Ca 2ϩ influx but suppressed peak Ca 2ϩ concentrations.
Acute CaMKII inhibition with a cell-permeable inhibitor peptide, AIP2, in dispersed wild-type C57/BL6 islet cell clusters also significantly affected Ca 2ϩ homeostasis (Fig. 5, C and D). In contrast to the effects of chronic CaMKII inhibition (see above), acute CaMKII inhibition using AIP2 increased basal Ca 2ϩ levels (2 mM glucose) and delayed the initiation of Ca 2ϩ influx following glucose stimulation (Fig. 5C). However, acute CaMKII inhibition with AIP2 suppressed total glucose-induced increases in Ca 2ϩ , similar to the effects of chronic CaMKII inhibition. Notably, Ca 2ϩ homeostasis in EAC3I-GFP ␤-cells treated with AIP2 was similar to EAC3I-GFP islet Ca 2ϩ without AIP2 (Fig. 5C), indicating that these acute changes in basal ␤-cell Ca 2ϩ are specifically caused by inhibition of CaMKII activity. The somewhat different effects of chronic in vivo and acute in vitro CaMKII inhibition on ␤-cell Ca 2ϩ homeostasis presumably reflect long term adaptive responses to CaMKII inhibition that may be influenced by the in vivo environment. Nevertheless, the peak glucose-stimulated increase in Ca 2ϩ concentration was significantly reduced by both acute and chronic CaMKII inhibition.
CaMKII Modulates ER Ca 2ϩ Storage in ␤-Cells-CaMKII modulates ER Ca 2ϩ uptake and release in cardiomyocytes (19 -23). Therefore, we investigated whether the alterations in basal Ca 2ϩ levels following ␤-cell CaMKII inhibition are due to changes in the intracellular Ca 2ϩ stores. ␤-Cell ER Ca 2ϩ levels were determined by inhibiting the sarco/endoplasmic reticulum ATPase (SERCA) with thapsigargin in the absence of extra-cellular Ca 2ϩ and measuring the resulting elevation of cytosolic Ca 2ϩ (Fig. 5, G and H). We found that ER Ca 2ϩ stores were significantly reduced in islets with chronic ␤-cell inhibition of CaMKII compared with controls (RIP-rtTA islets; Fig. 5, G and FIGURE 5. ␤-Cell CaMKII increases cytoplasmic and endoplasmic reticulum Ca 2؉ levels. A, relative islet Ca 2ϩ levels following glucose-stimulation (14 mM); islets (54 from RIP rtTA and 58 from RIP rtTA EAC3I) isolated from four independent animals. B, relative islet Ca 2ϩ influx during 6 min post initiation of glucose-stimulated islet Ca 2ϩ (IC) influx. C, relative islet Ca 2ϩ levels from islet cell clusters (ϳ20 cells per cluster) following glucose stimulation(14 mM); AIP2 (pretreated with 20 M cell permeable AIP2 for 20 min), vehicle (pretreated with vehicle for 20 min). D, relative islet Ca 2ϩ influx (AUC) during 6 min post initiation of glucose-stimulated islet Ca 2ϩ influx. E, relative islet Ca 2ϩ levels from GFP positive islet cells with (gray trace, RIP rtTA EAC3I-GFP) or without EAC3I (black trace, RIP rtTA tetO-GFP) in 5.6 mM glucose KRB in response to 20 mM KCl stimulation (black bar). F, relative KCl induced Ca 2ϩ increase (AUC) determined from ␤-cell Ca 2ϩ influx during KCl treatment. G, relative islet Ca 2ϩ levels during ER Ca 2ϩ release with 2 M thapsigargin (with no extracellular Ca 2ϩ ). H, relative ER Ca 2ϩ levels (AUC) determined from islet Ca 2ϩ influx during thapsigargin treatment. I, relative islet Ca 2ϩ levels during carbachol-induced (CCh) ER Ca 2ϩ release (with no extracellular Ca 2ϩ ). J, relative carbachol-induced ER Ca 2ϩ release (AUC) determined from islet Ca 2ϩ influx during carbachol treatment. Error bars represent means Ϯ S.E.; ***, p Ͻ 0.001.

H).
To confirm that the reduction in thapsigargin-stimulated ER Ca 2ϩ release was due to diminished ER Ca 2ϩ levels, islets were also stimulated with the insulin secretagogue carbachol in the absence of extracellular Ca 2ϩ to induce release of ER Ca 2ϩ via IP3 receptors. ER Ca 2ϩ release induced by carbachol treatment was found to be significantly reduced in islets with ␤-cell CaMKII inhibition (EAC3I-GFP) compared with control islets (RIP-rtTA) (Fig. 5, I and J). Therefore, CaMKII regulates ␤-cell cytoplasmic Ca 2ϩ at least in part through modulation of ER Ca 2ϩ handling.
CaMKII Activity Is Required for ␤-Cell L-type Ca 2ϩ Channel Facilitation-GSIS is driven by Ca 2ϩ influx via LTCCs. CaMKII facilitates voltage-induced Ca 2ϩ entry via LTCCs in myocytes and neurons by multiple mechanisms involving different LTCC subunits (24 -29). Therefore, we tested whether ␤-cell LTCCs also show voltage-induced facilitation that is dependent on CaMKII activity. Ca 2ϩ currents were measured in response to a 100-ms voltage step to 30 mV before and after a strong positive voltage prepulse for 200 ms (160 mV, Fig. 6, A and B). Control ␤-cells (RIP-rtTA-cells) show significant Ca 2ϩ channel facilitation following the positive voltage prepulse (current increase of 42.2 Ϯ 4.8%, n ϭ 19; Fig. 6, B and C), whereas ␤-cells with CaMKII inhibition show minimal voltage-induced Ca 2ϩ channel facilitation (current increase of 11.1 Ϯ 3.7%, n ϭ 25; Fig. 6, B and C). This indicates that CaMKII activity facilitates ␤-cell Ca 2ϩ channels following depolarization. To determine whether this change in LTCC facilitation influenced ␤-cell Ca 2ϩ currents, LTCC currents were recorded with voltage steps from Ϫ70 to ϩ70 mV in 10-mV increments. To prevent Ca 2ϩ induced changes in LTCC currents before the first recording, the whole-cell configuration was obtained with a constant holding potential of Ϫ80 mV in low glucose (2 mM) and only the first set of voltage step recordings for each cell was utilized for subsequent analysis. Interestingly, under these conditions, ␤-cells with CaMKII inhibition show significantly reduced Ca 2ϩ currents in response to voltage steps between 0 and ϩ40 mV when compared with control ␤-cells (RIP-rtTA-cells; Fig.  6, C and D). However, there is no change in VDCC activation or inactivation with CaMKII inhibition (Fig. 6, E and F). Thus, CaMKII enhances Ca 2ϩ entry into ␤-cells via LTCCs.
Glucose-stimulated activation of LTCCs is responsible for the upstroke of the action potential (AP), resulting in ␤-cell firing (30,31). Therefore, we measured mouse ␤-cell membrane potentials to test whether CaMKII modulation of LTCCs affects ␤-cell AP firing. The AP firing frequency after a 10-min exposure to 14 mM glucose was significantly reduced in islet ␤-cells expressing EAC3I-GFP (1.39 Ϯ 0.17 Hz) compared with control ␤-cells (RIP-rtTA-cells, 1.9 Ϯ 0.16 Hz) (Fig. 7, A, B, and  E). Furthermore, inhibition of ␤-cell K ATP channels by a 2-min exposure to tolbutamide increases the AP firing frequency in control ␤-cells (RIP-rtTA, 2.35 Ϯ 0.10 Hz), but this increase is significantly attenuated in EAC3I-GFP expressing ␤-cells (1.78 Ϯ 0.21 Hz) (Fig. 7, C-E). Thus, CaMKII has an important role in glucose-induced and LTCC-dependent increase in APfiring frequency, which further increases Ca 2ϩ influx.
CaMKII Enhances ␤-Cell Function in the Presence of Insulin Resistance-To determine the role of CaMKII in augmenting Ca 2ϩ influx and insulin secretion during periods of increased insulin demand, RIP-rtTA and EAC3I-GFP mice (n ϭ 5 each) were placed on a HFD at 3 weeks of age to induce systemic insulin resistance. GTTs revealed that both groups of mice exhibited comparable glucose tolerance, as expected because the eAC3I-GFP transgene is not expressed under these conditions (Fig. 8, A and C). Both cohorts of animals were then placed on doxycycline while the HFD was continued. Reassessment of glucose tolerance after 2 weeks on doxycycline revealed a substantial impairment of glucose tolerance in animals with EAC3I-GFP expression in ␤-cells when compared with RIP-rtTA animals (Fig. 8, B and C, n ϭ 5 per group). Because increased Ca 2ϩ influx augments insulin secretion during conditions of insulin resistance (32), glucose-stimulated Ca 2ϩ influx was also monitored in islets of HFD-treated animals. Glucose-stimulated Ca 2ϩ influx was significantly diminished in EAC3I-GFP expressing islets compared with RIP-rtTA islets, which were both isolated from HFD-treated mice (Fig. 8, D and  E). This demonstrates an important role for CaMKII in augmenting glucose-stimulated ␤-cell Ca 2ϩ influx and insulin secretion under conditions of systemic insulin resistance.

DISCUSSION
It is well established that Ca 2ϩ entry into pancreatic ␤-cells stimulates insulin secretion and that CaMKII activity influences Ca 2ϩ homeostasis in many cell types (20,(33)(34)(35). Although CaMKII has been implicated in regulating islet insulin secretion, mechanisms underlying CaMKII action in ␤-cells are poorly understood (2). The results presented here demonstrate that CaMKII modulates Ca 2ϩ handling and insulin secretion in a mouse model with conditional ␤-cell expression of a CaMKII inhibitory peptide. The results suggest that CaMKII enhances glucose tolerance by amplifying glucose-stimulated Ca 2ϩ entry and insulin secretion.
Ca 2ϩ -dependent amplification and/or reduction of the channels required for cytosolic Ca 2ϩ entry dynamically modulates Ca 2ϩ signals in many cell types. Amplification can occur through many mechanisms, including Ca 2ϩ -dependent activation of ER Ca 2ϩ release via ryanodine receptor (RYR) channels, as well as through Ca 2ϩ -dependent facilitation of plasma membrane LTCCs (33,36,37). Although some of these mechanisms involve direct binding of Ca 2ϩ -binding proteins (e.g. calmodulin) to the channels, others involve additional Ca 2ϩ -activated proteins such as CaMKII (33). In cardiac cells, CaMKII regulates multiple proteins involved in Ca 2ϩ homeostasis, tightly modulating excitation-contraction and excitation-transcription coupling (33,36,37). Similarly, we find that CaMKII controls Ca 2ϩ homeostasis in pancreatic ␤-cell, in this case, modulating excitation-secretion coupling. When activated during glucose-stimulation, ␤-cell CaMKII facilitates LTCCs and amplifies glucose-stimulated islet Ca 2ϩ entry. Although LTCC facilitation by CaMKII can be induced by depolarization, it has also been shown to result from ER Ca 2ϩ release in neurons as well as sarcoplasmic reticulum Ca 2ϩ release in myocytes (26,38,39). Therefore, the CaV1.2 and/or CaV1.3 LTCCs in ␤-cells may undergo CaMKII-dependent facilitation via multiple mechanisms under different physiological conditions, which influence insulin secretion. The results suggest that ␤-cell CaMKII provides a Ca 2ϩ -sensitive feedback loop for enhancing GSIS secretion by amplifying Ca 2ϩ influx.
CaMKII can also indirectly regulate LTCCs by regulating the plasma membrane potential via K ϩ channel modulation (40). Data presented here suggest that CaMKII does not modulate ␤-cell K V and K ATP currents. We also find that CaMKII enhances AP firing frequency during K ATP inhibition. Our data suggest that CaMKII regulates ␤-cell electrical activity and Ca 2ϩ entry independently of any effects on K ATP or K V channel activity. However, CaMKII can inhibit the inward rectifier subunit of the ␤-cell K ATP channel complex (Kir6.2) (41). Moreover, CaMKII modulates H 2 O 2 activation of sarcolemmal K ATP channels and oxidative stress activates K V 2.1 channels (42,43). Thus, it is important to note that our data do not exclude potential effects of CaMKII on ␤-cell K ϩ channels under specific conditions not assessed in this study such as oxidative stress.
Endoplasmic reticulum Ca 2ϩ stores also play an important role in regulating ␤-cell excitability and GSIS (44 -47). CaMKII modulates ER Ca 2ϩ levels by regulating both RYR and SERCA (21)(22)(23). RYR2 is phosphorylated by CaMKII in islets and mutations of RYR2 that mimic CaMKII phosphorylation increase ␤-cell ER Ca 2ϩ release (23). Therefore, diminished ER Ca 2ϩ leak through RYRs following chronic inhibition of ␤-cell CaMKII may cause the reductions in both thapsigargin-induced ER Ca 2ϩ release and basal cytoplasmic Ca 2ϩ levels. However, carbachol-induced ER Ca 2ϩ release through IP3 receptors is also decreased in islets with ␤-cell CaMKII inhibition. Therefore, ␤-cell ER Ca 2ϩ levels appear to be reduced by CaMKII inhibition, possibly due to SERCA inhibition. A brief reduction in SERCA activity with acute inhibition of CaMKII would account for the increase of basal Ca 2ϩ (Fig. 5C) due to the inability of ER to take up cytoplasmic Ca 2ϩ . A sustained reduction of SERCA activity with chronic CaMKII inhibition would ultimately decrease ␤-cell ER Ca 2ϩ stores (Fig. 5A). The exact mechanism(s) responsible for CaMKII modulation of ␤-cell ER Ca 2ϩ handling are currently being investigated.
CaMKII-dependent modulation of Ca 2ϩ influx likely causes the changes in GSIS reported here because ionomycin-stimulated insulin secretion is unaffected by CaMKII inhibition. However, other studies suggest roles for CaMKII in regulating components of insulin granule priming and or fusion. For example, one study found that CaMKII inhibition decreased insulin secretion even when islet intracellular calcium levels are clamped (48). These effects may involve the direct phosphorylation of synapsin 1 and microtubule-associated protein, MAP2, by CaMKII (49,50). The studies presented here confirm that synapsin1 phosphorylation is indeed regulated by CaMKII. Synapsin1 controls the targeting of synaptic-like vesicles in primary pancreatic ␤-cells and rat insulinoma cells (17) and is localized to insulin granules in mouse insulinoma cells (51). This indicates the possibility that ␤-cell synapsin1 either directly influences insulin granule release via association with the insulin granule or indirectly influences insulin secretion FIGURE 8. ␤-Cell CaMKII protects animals from glucose intolerance following exposure to a high fat diet by enhancing glucose stimulated Ca 2؉ entry.
A, glucose tolerance test on mice (black line, RIP rtTA; gray line, RIP rtTA EAC3I) following 2 months on a high fat diet. B, glucose tolerance test on the mice shown in A 2 weeks post doxycycline (Dox.) treatment (gray line, ␤-cell CaMKII inhibition; black line, controls) following 2.5 months on a high fat diet. C, area under the curve data for the GTT of RIPrtTA (black bars) and EAC3I-GFP (gray bars) mice before and after doxycycline treatment. D, relative islet Ca 2ϩ levels following glucose-stimulation (14 mM); islets (n Ͼ 70 islets from each group, RIP rtTA and RIP rtTA EAC3I) isolated from four independent animals. E, relative islet Ca 2ϩ influx (AUC) during 6 min following the initiation of islet glucose-stimulated Ca 2ϩ entry. Error bars represent means Ϯ S.E.; *, p Ͻ 0.05. through release of neurotransmitters from synaptic like vesicles. MAP2 is also phosphorylated by CaMKII and may influence insulin granule priming through insulin granule trafficking to the membrane on microtubules. This would be predicted to influence second phase insulin secretion, and interestingly, the studies presented here identify significant reductions in second phase insulin secretion when CaMKII is inhibited in islet ␤-cells. Thus, MAP2 may be important in regulating trafficking of insulin granules in response to CaMKII activation during GSIS. Future studies will determine the roles of synapsin 1 and MAP2 regulation of insulin secretion and islet function in response to glucose activation of CaMKII. In summary, although the changes in Ca 2ϩ induced by ␤-cell CaMKII inhibition are likely to impair first phase insulin secretion, our data cannot exclude additional roles for CaMKII during insulin secretion such as granule trafficking and priming during second phase insulin secretion.
Although CaMKII physiologically modulates islet function, not much is known about the influence of CaMKII signaling during the pathogenesis of diabetes (2). Perturbations in CaMKII signaling are observed following palmitate treatment of insulinoma ␤-cells (52). Therefore, changes in CaMKII signaling may contribute to aberrant ␤-cell function following HFD feeding, during the progression of diabetes (52,53). Indeed, we found that the inhibition of CaMKII in mice maintained on a HFD results in a rapid exacerbation of glucose intolerance. These data indicate that ␤-cell CaMKII activity is critical for the maintenance of glucose-stimulated islet Ca 2ϩ influx and thus insulin secretion in the face of chronic hyperglycemia and the initial progression of insulin resistance induced by a high fat diet, thereby protecting mice from glucose intolerance. This indicates the exciting possibility that augmenting CaMKII activity may also increase human glucose-stimulated ␤-cell Ca 2ϩ entry and enhance GSIS in a diabetic setting. Future studies will determine how CaMKII signaling influences human ␤-cell GSIS during the pathogenesis of diabetes.
In conclusion, this study suggests that CaMKII plays an important role as a ␤-cell Ca 2ϩ sensor that modulates Ca 2ϩ handling under both high energy/glucose and low energy/glucose conditions. CaMKII amplifies Ca 2ϩ influx and insulin secretion during glucose stimulation and maintains cytoplasmic Ca 2ϩ levels under fasting conditions. Thus, CaMKII is a key Ca 2ϩ sensor that controls pancreatic ␤-cell Ca 2ϩ homeostasis to dynamically modulate insulin secretion.