Loss of α₂δ-1 Calcium Channel Subunit Function Increases the Susceptibility for Diabetes

Abstract

Reduced pancreatic β-cell function or mass is the critical problem in developing diabetes. Insulin release from β-cells depends on Ca²⁺ influx through high voltage– gated Ca²⁺ channels (HVCCs). Ca²⁺ influx also regulates insulin synthesis and insulin granule priming and contributes to β-cell electrical activity. The HVCCs are multisubunit protein complexes composed of a pore-forming α1 and auxiliary β and α₂δ subunits. α₂δ is a key regulator of membrane incorporation and function of HVCCs. Here we show that genetic deletion of α₂δ-1, the dominant α₂δ subunit in pancreatic islets, results in glucose intolerance and diabetes without affecting insulin sensitivity. Lack of the α₂δ-1 subunit reduces the Ca²⁺ currents through all HVCC isoforms expressed in β-cells equally in male and female mice. The reduced Ca²⁺ influx alters the kinetics and amplitude of the global Ca²⁺ response to glucose in pancreatic islets and significantly reduces insulin release in both sexes. The progression of diabetes in males is aggravated by a selective loss of β-cell mass, while a stronger basal insulin release alleviates the diabetes symptoms in most α₂δ-1^−/− female mice. Together, these findings demonstrate that the loss of the Ca²⁺ channel α₂δ-1 subunit function increases the susceptibility for developing diabetes in a sex-dependent manner.

In response to elevated blood glucose levels, pancreatic β-cells release insulin by a Ca²⁺ channel–dependent process, with a rapidly rising transient first phase followed by a sustained second release phase (1). Pharmacological block or genetic deletion demonstrated that Ca_V1.2 channels are responsible for insulin release in the first and second phase (2–4). Ca_V1.2 Ca²⁺ currents are also critical for shaping the electrical activity of pancreatic β-cells, and in the absence of Ca_V1.2 channels, the fast phase of insulin release is strongly delayed (2,5). Conversely, gain-of-function mutations in Ca_V1.2 channels lead to hypoglycemia and death (6). Genetic ablation of Ca_V1.3 channels has been shown to reduce the postnatal generation and survival of mouse β-cells (7), and human genetic polymorphisms reducing Ca_V1.3 expression increase the susceptibility for type 2 diabetes (8). R-type Ca²⁺ currents through Ca_V2.3 channels have been proposed to be necessary for the priming of insulin granules at the plasma membrane and thus for sustaining the tonic second-phase insulin release (9,10). Therefore, the tight control of Ca²⁺ entry through multiple high voltage–gated Ca²⁺ channels (HVCCs) is of critical importance for normal β-cell function and for the maintenance of β-cell mass in pancreatic islets. Any alteration in the composition and function of HVCCs is expected to lead to β-cell dysfunction and diabetes (11,12).

HVCCs are heteromeric membrane proteins composed of a pore-forming α₁-subunit, an intracellular β-subunit, and an extracellular α₂δ-subunit (13,14). The role of α₂δ-1 in the channel complex can vary, depending on the tissue in which it is expressed and on the specific α₁-isoform with which it associates. In neurons, the primary role of α₂δ-1 is to promote the plasma membrane incorporation of Ca_V1 and Ca_V2 channels (15–18), whereas in muscle cells, the α₂δ-1 subunit determines the characteristic current kinetics of Ca_V1.1 and Ca_V1.2 channels (19–21). Recent transcriptome analysis indicated that α₂δ-1 is the main α₂δ-isoform expressed in both human and mouse pancreatic islets (22). Several loss-of-function mutations in the CACNA2D1 gene have been identified that associate with epilepsy and intellectual disability (23), short QT syndromes (24,25), and cardiac arrhythmias (26). However, the physiological role of α₂δ-1 in endocrine cells is unknown and its importance for HVCC channel function in β-cells, on dynamic insulin release and glucose tolerance, has not been analyzed.

Here, we show that the constitutive knockout (KO) of the α₂δ-1 Ca²⁺ channel subunit (27) results in reduced first- and second-phase insulin release, in both male and female mice, without affecting the insulin sensitivity. A stronger basal insulin release partially alleviates the diabetes symptoms in most α₂δ-1^−/− females. However in α₂δ-1^−/− males and to a much lower extent in α₂δ-1^−/− females, a selective and progressive loss in β-cell mass, secondary to diabetes, amplifies the symptoms. Collectively, our data show that impaired α₂δ-1 function leads to diabetes in males and increases diabetes susceptibility in females. This critical role of the Ca²⁺ channel α₂δ-1 subunit in normal blood sugar regulation should prompt clinical studies investigating glucose tolerance in patients carrying CACNA2D1 loss-of-function mutations (23–26).

Research Design and Methods

α₂δ-1^−/− Mice

α₂δ-1^−/− mouse strain was previously generated and characterized (27,28). In the previous reports, the α₂δ-1^−/− mice were bred in a C57BL/6 genetic background, whereas our colony has a mixed C57BL/6 × 129J background. α₂δ-1^+/+ and α₂δ-1^−/− littermates were bred from heterozygous mice. All experiments were performed on 4- to 5-week-old animals unless stated otherwise. All animal experiments were performed in conformity with international laws and were approved by the Austrian Ministry of Science (BMWFW-66.008/0015-WF/V/3b/2014).

Immunohistochemistry

Males

After dissection, the pancreas was fixed overnight in buffered 4% paraformaldehyde and subsequently embedded in paraffin. Serial 4-μm-thick sections were stained with hematoxylin and eosin, for insulin (A0564, dilution 1:100; DAKO) and for glucagon (PA5-32424, dilution 1:50; Thermo Fisher Scientific). Islet numbers were determined by counting all islets in every sixth section of the whole pancreas. The islet size was calculated by integrating the area of at least 20 islets per mouse.

Females

Pancreas was perfused with 4% paraformaldehyde, stored overnight in 4% paraformaldehyde at 4°C, and subsequently prepared for cryopreservation: 30% sucrose in PBS (overnight, 4°C), 30% OCT (2 h, 4°C, Tissue-Tek; Sakura Finetek Europe B.V.), 100% OCT (1 h, 4°C), −20°C in the cryostat. Sections were cut 20 μm thick and stained for insulin (A0564 1/1,000 + Alexa568 [DAKO] and Ab175714, 1/4,000 [Abcam]) and glucagon (G2654, 1/200 + Alexa488 [Sigma-Aldrich] and A11029, 1/2,000 [Thermo Fisher Scientific]). Islet and cells numbers were counted in every section.

Isolation of Pancreatic Islets and β-Cells

Pancreatic islets were enzymatically isolated as previously reported (9). For Ca²⁺ fluorometry and insulin release, islets were cultured at 37°C for 2 h in RPMI1640 medium supplemented with 10% FCS, 2 mmol/L l-glutamine, 5 mmol/L glucose, 100 μg/mL streptomycin, and 100 IU/mL penicillin. To disperse the pancreatic cells, the islets were incubated in solution containing (mmol/L) 138 NaCl, 6 KCl, 3 MgCl₂, 5 HEPES, 3 glucose, 1 EGTA, and 1 mg/mL BSA for 10 min at 37°C and then mechanically dissociated. Single cells were cultured for at least 3 h in RPMI1640 medium before the start of the experiment.

Electrophysiology

Ionic currents were recorded from dispersed pancreatic islet cells using perforated-patch technique at room temperature using the Axopatch 200B amplifier. To record Ca²⁺ currents, the pipette solution contained (mmol/L) 76 CsSO₄, 10 CsCl, 10 KCl, 1 MgCl₂, and 5 HEPES with 120 μg/mL amphotericin B (Sigma-Aldrich), and the extracellular solution contained (mmol/L) 140 TEA-Cl, 5.6 KCl, 1.2 MgCl₂, 5 HEPES, 2.6 CaCl₂, and 5 glucose. To pharmacologically dissect the contribution of different HVCC isoforms to total β-cell Ca²⁺ currents, we used isradipine (2 μmol/L), a Ca_V1.2 and Ca_V1.3 L-type Ca²⁺ channel blocker, and SNX-482 (100 nmol/L), a Ca_V2.3 R-type channel blocker. For recording sodium currents and capacitance increase, the extracellular solution contained also (mmol/L) 138 NaCl and 20 TEA-Cl. The voltage dependence of the sodium current inactivation was used to identify β-cells (29). The increase in cell capacitance after step depolarizations was recorded at 32°C using the lock-in sine+DC mode of HEKA amplifier controlled by the PULSE software.

Ca²⁺ Fluorometry

Isolated pancreatic islets were loaded with Indo-1 Ca²⁺ indicator (6 μmol/L; Invitrogen) for 1 h at room temperature, followed by 30 min at 37°C. Single islets were perfused with 37°C solution ([mmol/L] 140 NaCl, 0.5 NaH₂PO₄, 0.5 MgSO₄-7H₂O, 2.5 CaCl₂, 2 NaHCO₃, 5 HEPES, 3.6 KCl) with a perfusion rate of 2 mL/min. Fluorescent signals (400 and 480 nm) were recorded using two photomultipliers (PTI, South Brunswick, NJ).

Dynamic Insulin Release and Insulin Content

For each animal, a group of 20 islets of similar size were perfused with the same solution as for Ca²⁺ imaging at 250 μL/min flow rate, and the insulin concentration of the 2-min fractions was measured with either a normal sensitive (α₂δ-1^+/+ mice) or an ultrasensitive (α₂δ-1^−/− mice) ELISA kit (Mercodia). The insulin content was extracted from groups of 10 islets using the acid-ethanol method (2,9).

Quantitative TaqMan PCR

Total RNA was extracted from isolated islets using the RNeasy Mini kit (Qiagen) and reverse transcribed (SuperScriptII; Invitrogen). The relative abundance of α₂δ isoforms was assessed by TaqMan quantitative PCR (50 cycles) using a standard curve based on PCR products of known concentrations and normalization to endogenous controls (30,31). TaqMan gene expression assays were purchased from Applied Biosystems (31).

Statistical Analysis

Data analysis was performed using Clampfit 10.2 (Axon Instruments), Sigma Plot 11 (Systat Software, Inc.), Excel 11 (Microsoft), or GraphPad Prism 7. All values are presented as mean ± SE for the indicated number of experiments (n), except if stated otherwise. Statistical significance was calculated using unpaired Student t test (Welch test for differing variances) or one-way ANOVA followed by Bonferroni post hoc test as indicated.

Results

All α₂δ-1^−/− Males and Some α₂δ-1^−/− Female Mice Have Altered Metabolic Parameters and Shorter Life Span

The original characterization of the α₂δ-1^−/− mouse reported decreased basal myocardial contractility and relaxation due to altered Ca_V1.2 Ca²⁺ influx. Otherwise, α₂δ-1^−/− mice appeared normal, except that KO males displayed a tendency for bladder dilatation (27) (A. Schwartz, personal communication). Indeed, anatomical reexamination showed severely enlarged bladders and kidneys, especially in KO males >2 months of age (Fig. 1A). These symptoms were accompanied by a dramatic increase in water consumption and urine excretion (Fig. 1B and C, solid lines). In contrast, most KO females showed no signs of bladder extension, increased water consumption, or urine excretion. However, when affected (one out of five females examined in metabolic cages), the severity of the symptoms in KO females was comparable to that of the males (Fig. 1B, red dotted line).

Figure 1.

Phenotypic characterization of α₂δ-1^−/− mice. A: Adult α₂δ-1^−/− males are lean and have enlarged bladder (arrows) and kidneys compared with their WT littermates (scale bar, 5 mm). Water consumption (B) and urine production (C) measured using metabolic cage (Tecniplast, Milan, Italy) (n = 5 of each sex and genotype). The dashed red line represents an example of a symptomatic female. D: Genotype of the newborn mice of heterozygous (HT) parents (n > 450). E: Postnatal survival rate of WT and α₂δ-1^−/− mice (n > 40 of each sex and genotype).

Deletion of α₂δ-1 did not affect the survival rate in utero, as the genotype of newborn mice of heterozygous parents followed a Mendelian ratio (Fig. 1D). However, postnatal survival of KO males declined rapidly (Fig. 1E), with <50% of mice surviving at 5 months of age (16 out of 41, survival rate 39%). In contrast, the survival rate of the KO females (93%) was hardly affected, as out of 41 females, only 3 were symptomatic and died within 1 year. Together these symptoms strongly indicated that a progressive diabetic phenotype reduced survival rate, especially in males.

α₂δ-1^−/− Mice Are Glucose Intolerant But Insulin Sensitive

Indeed, KO males on average had approximately threefold higher basal blood glucose levels at 1 month, increasing to approximately ninefold at 5 months of age (Fig. 2A). Young KO females (1 month of age) had modestly elevated basal blood glucose levels compared with their wild-type (WT) littermates, but glucose levels normalized as the mice matured (Fig. 2A). Intraperitoneal glucose tolerance test (IPGTT) in 6 h–fasted mice clearly demonstrated abnormal glucose tolerance for both KO males and females, although to a different extent. In WT males, blood glucose levels increased by approximately twofold 15 min after glucose injection (1g/kg, i.p.) and then returned to fasting levels within 1 h. In KO males, the same dose of glucose raised blood glucose levels to values approximately three times higher than in WT mice and their blood glucose levels did not normalize even after 2 h. α₂δ-1 deletion also strongly reduced glucose tolerance in females. However, both WT and KO females cleared glucose significantly faster than the corresponding males (Fig. 2B).

Figure 2.

α₂δ-1^−/− mice are glucose intolerant but insulin sensitive. A: Nonfasting blood glucose levels in WT and α₂δ-1^−/− mice (n = 5 of each sex and genotype). B: IPGTT using 1 mg/g body weight glucose (males WT n = 6, α₂δ-1^−/− n = 8, females both genotypes n = 7). C: Intraperitoneal insulin sensitivity test after bovine insulin injection (0.75 units/kg body weight) (WT male n = 4, α₂δ-1^−/− male n = 9, WT female n = 10, α₂δ-1^−/− female n = 7). D: Quantitative real-time PCR analysis of α₂δ-1, -2, and -3 mRNA expression in isolated islets (WT n = 3, α₂δ-1^−/− n = 2). All values are means ± SEM. Statistics with two-way ANOVA. **P < 0.01, ***P < 0.001, comparison between WT and KO mice of the same sex; ##P < 0.01, ###P < 0.001, comparison between male and female KO mice.

Because skeletal muscle, the primary tissue regarding glucose utilization, expresses high levels of α₂δ-1, we next examined insulin sensitivity in WT and KO mice after 6 h fasting. Within 30 min after injection of insulin (0.75 units/kg, i.p.), plasma glucose levels were lowered to the same concentration in KO and WT mice (Fig. 2C). This indicated that general KO of α₂δ-1 did not result in reduced insulin sensitivity, pointing to pancreatic islet dysfunction as a cause for the glucose intolerance in α₂δ-1^−/− mice. This possibility was substantiated by our quantitative qRT-PCR data, which show that α₂δ-1 is by far the dominant α₂δ isoform in pancreatic islets and that genetic ablation of α₂δ-1 is not compensated by a comparable upregulation of α₂δ-2 or -3 transcripts (Fig. 2D).

α₂δ-1^−/− Males Show a Progressive Reduction in β-Cell Mass

To quantify possible effects of α₂δ-1 KO on the number, size, and composition of pancreatic islets, we immunolabeled pancreas sections from 6- and 14-week-old males and 15–18-week-old females with antibodies against insulin and glucagon. In 6-week-old WT and KO males, the insulin-positive β-cells occupied the core of the pancreatic islet whereas glucagon-positive a-cells surrounded them in the periphery (Fig. 3A). At 14 weeks of age, the α₂δ-1^−/− male islets had fewer insulin-positive β-cells (Fig. 3B). Quantification of the islet density revealed a small nonsignificant reduction at 6 weeks (Fig. 3C), and their average size was normal (Fig. 3D). However, in 14-week-old KO mice, the islet density was significantly reduced by 66% (Fig. 3C), and this was accompanied by a 46% reduction in islet size and almost complete loss of larger islet profiles (Fig. 3D).

Figure 3.

α₂δ-1^−/− deletion induces a progressive and selective loss of β-cells in males. A: Insulin and glucagon staining in fixed pancreatic sections from 6-week-old males (n = 3 both genotypes). B: Same as in A but at 14 weeks of age (WT n = 5; α₂δ-1^−/− n = 3). C: Box plot analysis of the islet density in pancreatic sections from 6- and 14-week-old males. D: Dot plot indicating the islet size in the same pancreatic sections. Box plot depicting the β-cell density (E) and α-cell density (F) in islets of 6- and 14-week-old males. G: Insulin and glucagon staining in fixed pancreatic sections from 15- to 18-week-old females (WT n = 2, α₂δ-1^−/− n = 3). H: Dot plot indicating the islet size in the pancreatic sections from females. Box plot depicting the β-cell density (I) and α-cell density (J) in female islets. All values are means ± SEM. Statistics with Mann-Whitney nonparametric test: **P < 0.01; ***P < 0.001.

To examine whether the reduction in the islet size is caused by specific loss of β-cells, we quantified the density of α- and β-cells (Fig. 3E and F). Whereas the β-cell density was significantly decreased in KO compared with WT islets at 6 and 14 weeks of age, the α-cell density did not decline. Thus, whereas severe hyperglycemia was observed in 4-week-old α₂δ-1^−/− males (compared with Fig. 2), pancreatic islets initially developed normally. To examine whether deletion of α₂δ-1 alters the β-cell mass in the absence of severe diabetes symptoms, we performed a similar analysis in nonsymptomatic 15–18-week-old females. In α₂δ-1^−/− females, the structure of the islet was intact (Fig. 3G), whereas the surface area of the islet was rather increased by 24% (Fig. 3H). The β-cell density was decreased by 9.7% (Fig. 3I), but the α-cell numbers increased by 42% in α₂δ-1^−/− islets compared to WT (Fig. 3J).

Knockout of α₂δ-1 Reduces Ca²⁺ Influx in Pancreatic β-Cells Equally in Females and Males

To identify the molecular mechanism responsible for hyperglycemia, we first analyzed the properties of the HVCC Ca²⁺ currents in dissociated pancreatic β-cells (9). Whole-cell patch-clamp recordings revealed greatly diminished Ca²⁺ currents in male and female α₂δ-1^−/− mice (Fig. 4A and G). The I/V curves (current-voltage relationship) showed an ~70% reduced mean maximal current density in both KO males and females (Fig. 4B and H), whereas the voltage dependence of activation was not significantly changed (Fig. 4C and I). Furthermore, the Ca²⁺ currents recorded from α₂δ-1^−/− β-cells displayed altered kinetic properties (Fig. 4A and G) with a significantly slowed activation time course (time to peak) in males (from 5.2 ± 0.2 to 8.4 ± 1.0 ms) and females (5.9 ± 0.3 to 8.8 ± 1.1 ms) (Fig. 4D and J). The fractional inactivation after 200 ms was also significantly reduced in both males (from 63.1 ± 2.3% to 45 ± 5.1%) (Fig. 4F) and females (from 63.2 ± 2.7% to 46.3 ± 4.2%) (Fig. 4L), but not after 50 ms (Fig. 4E and K). Therefore, the lack of α₂δ-1 equally affects Ca²⁺ influx in β-cells of male and female KO mice.

Figure 4.

β-Cell Ca²⁺ current density and kinetics are equally affected by α₂δ-1 deletion in both male and female mice. Representative Ca²⁺ current traces (A and G) and corresponding I/V curves (B and H) recorded in isolated β-cells of male (top) and female mice (bottom) (α₂δ-1^−/− male and females n = 13, WT male n = 22, WT female n = 16). C and I: Voltage dependence of Ca²⁺ conductance. D and J: Activation kinetics (time to peak) of Ca²⁺ current. Fractional current inactivation after 50-ms (E and K) and 200-ms (F and L) depolarization. Ca²⁺ currents were elicited using 200-ms depolarization steps from −80 to +60 mV in 10-mV increments and analyzed as previously described (20). All values are means ± SEM. Statistics with Student t test: **P < 0.01; ***P < 0.001.

All HVCC Types Are Equally Reduced in α₂δ-1^−/− β-Cells

Several different HVCC isoforms contribute to β-cell function (11,32). Application of the L-type Ca²⁺ channel blocker isradipine (2 μmol/L) reduced the β-cell current density by ~70% (Fig. 5A–C). Additional application of the Ca_V2.3 channel blocker SNX-482 (100 nmol/L) further reduced the current density by 15% of total. The remaining ~15% current corresponds to P/Q- and N-type (Ca_V2) channels (2). The smaller Ca²⁺ currents in α₂δ-1^−/− β-cells were blocked by isradipine (~70%) and SNX-482 (~15%) to similar extents (Fig. 5D–F). This shows that the genetic deletion of α₂δ-1 does not alter the fractional contribution of L- and R-type currents in β-cells.

Figure 5.

Deletion of α₂δ-1 equally affects all β-cell HVCC isoforms. Sample recording of Ca²⁺ currents in WT male (A) and α₂δ-1^−/− male (D) β-cells without drug application and after perfusion with 2 μmol/L isradipine or 2 μmol/L isradipine + 100 nmol/L SNX-482. I/V curve (B and E) and stack column (C and F) representation of the current density in the presence or absence of specific Ca²⁺ channel blockers. Ca²⁺ currents were elicited using 200-ms depolarization steps from −80 to +60 mV in 10-mV increments and analyzed as previously described (20). All values are means ± SEM. Statistics with Student t test.

Knockout of α₂δ-1 Differentially Alters the Ca²⁺ Dynamics in Pancreatic Islets of Male and Female Mice

How do reduced current density and slowed inactivation of Ca²⁺ currents alter glucose-induced Ca²⁺ signals in the α₂δ-1^−/− pancreatic islets? To answer this question, we loaded isolated pancreatic islets with the ratiometric Ca²⁺ indicator Indo-1 (6 μmol/L) and recorded the fluorescent signal from whole islets (9). In 5 mmol/L extracellular glucose, the pancreatic islets showed stable resting fluorescence levels. One minute after increasing glucose to 15 mmol/L, a strong rise in cytoplasmic free Ca²⁺ concentration ([Ca²⁺]i) was recorded in all islets (Fig. 6A and E). However, the time to peak of the first-phase Ca²⁺ response was more than doubled in α₂δ-1^−/− islets of both sexes compared with controls (Fig. 6B and F). Moreover, in male α₂δ-1^−/− islets, the magnitude of the glucose-induced Ca²⁺ transient was significantly smaller compared with the WT islets (Fig. 6C and D).

Figure 6.

Ca²⁺ transients are slower and have smaller amplitude in α₂δ-1^−/− islets. Average Ca²⁺ transients in single islets isolated from male (A) and female (E) mice (male WT n = 16, α₂δ-1^−/− n = 15, female WT n = 14, α₂δ-1^−/− n = 10). The time to peak of the first phase in male islets (B) and female islets (F). The peak amplitudes (C) and integrals (D) of the glucose-dependent and glucose-independent Ca²⁺ transients in male islets. Amplitudes (G) and integrals (H) of the glucose-dependent and glucose-independent transients in female islets. I: Average [Ca²⁺]i transients in single isolated male islets from WT (n = 20, four mice) and α₂δ-1^−/− (n = 26, three mice) stimulated with 50 mmol/L KCl or 1 μmol/L carbachol (CCh). J: The average peak amplitude of the KCl and carbachol-induced [Ca²⁺]i transients. All values are means ± SEM. Statistics with Student t test: *P < 0.05; **P < 0.01; ***P < 0.001. A.U., arbitrary units.

To bypass possible alterations in glucose metabolism or excitability of α₂δ-1^−/− β-cells, we applied the K_ATP channel blocker tolbutamide (100 μmol/L) (33,34). The [Ca²⁺]i transient in response to tolbutamide was also substantially decreased in male α₂δ-1^−/− islets compared with WT controls. Whereas the female α₂δ-1^−/− islets displayed significantly slowed activation of the first-phase glucose response, the amplitude of the glucose- or tolbutamide-induced [Ca²⁺]i transient increase was not reduced (Fig. 6G and H). Passive depolarization with 50 mmol/L KCl induced a significantly lower [Ca²⁺]i peak in α₂δ-1^−/− islets of both sexes compared with WT controls. To investigate whether altered intracellular Ca²⁺ store load contributes to reduced islet [Ca²⁺]i peak in male islets, we stimulated the Ca²⁺ release via InsP3R (inositol triphosphate receptor) with 1 μmmol/L carbachol (Fig. 6I and J). Whereas the amplitude of the KCl step was significantly reduced in α₂δ-1^−/− islets (WT 0.76 6 0.03, KO 0.37 6 0.01), the amplitude of the carbachol-induced [Ca²⁺]i transient was similar for both WT (0.37 6 0.02, n = 20) and α₂δ-1^−/− islets (0.38 6 0.02, n = 26). Taken together, the [Ca²⁺]i recordings in isolated islets demonstrate that the loss of α₂δ-1 function results in a severely delayed [Ca²⁺]i response and also in males a smaller amplitude and total integrated [Ca²⁺]i responses. The reduced [Ca²⁺]i transient is caused by reduced Ca²⁺ influx and not by altered intracellular store release.

α₂δ-1 Deletion Drastically Reduces Insulin Release but Not Insulin Content of Pancreatic Islets

Because in α₂δ-1^−/− mice the Ca²⁺ influx through L-type and R-type HVCCs is dramatically reduced (Figs. 4 and 5) and the first-phase islet [Ca²⁺]i transient is substantially delayed (Fig. 6), we expected a change in the dynamics of insulin release. To measure this, we perfused 20 isolated islets and collected fractions every 2 min for analysis of insulin concentration. As expected, after increasing from 5 to 15 mmol/L glucose, WT islets responded with a fast peak of insulin release followed by a smaller sustained release phase. At the end of the experiment, massive insulin release was induced by passively depolarizing the islets with KCl (50 mmol/L). In WT females, but not in males, the second-phase insulin release steadily increased to levels above the initial peak (Fig. 6A and E). The distinct phases of insulin release were observed in both WT and α₂δ-1^−/− islets. However, α₂δ-1 deletion substantially reduced all phases of insulin release in islets of both sexes. In α₂δ-1^−/− male islets, the maximal amplitude of the first phase was reduced by 78% (Fig. 6B) and the integral of the second phase by 77% (Fig. 6C). Peak KCl-induced release was reduced by 51%. In female islets, the amplitude of the first phase was reduced by 60%, integrated insulin secreted during the slow phase by 77%, and the peak of KCl-induced insulin release by 61% (Fig. 6F and G). Notably, these differences were not due to different insulin content in islets from WT and KO mice (Fig. 7D and H). To probe whether the reduced islet insulin release is caused by impaired β-cell secretory function, we quantified the increase in membrane capacitance after 10 consecutive 500-ms depolarization steps in male β-cells (Fig. 7I). In α₂δ-1^−/− β-cells, the average capacitance at the end of the stimulus train was 179.9 ± 17.9 fF (n = 24) and in WT cells 519.5 ± 52.3 fF (n = 25) (Fig. 7J). The increase in membrane capacitance after the first depolarization step, reflecting the ready releasable pool of insulin vesicles, was significantly smaller in α₂δ-1^−/− β-cells (17.6 ± 6.6 fF) compared with WT β-cells (64.9 ± 17.9 fF) (Fig. 7K). The average increase in membrane capacitance induced by each subsequent depolarization step was also significantly reduced as the α₂δ-1^−/− β-cell capacitance increased by only 22.6 ± 1.9 fF and that of WT β-cells by 51.1 ± 4.9 fF (Fig. 7L).

Figure 7.

Deletion of α₂δ-1^−/− reduces both phases of insulin release. Average insulin secretion from male islets (WT and α₂δ-1^−/− n = 6) (A) and female islets (WT n = 6, α₂δ-1^−/− n = 7) (E). The peak of the first-phase glucose-induced and of KCl-triggered insulin secretion from male (B) and female (F) islets. Integral of the insulin secretion during the different phases from male (C) and female (G) islets. Islet insulin content in males (WT n = 6, α₂δ-1^−/− n = 7) (D) and females (WT n = 6, α₂δ-1^−/− n = 10) (H). I: Average exocytosis evoked by a train of 10 depolarization pulses, 500 ms long, from 270 to 0 mV in WT (n = 25, three mice) and α₂δ-1^−/− (n = 24, three mice). J: Average capacitance increase after 10 depolarization pulses. K: Capacitance increase after the first depolarization pulse. L: Mean capacitance increase with each depolarization pulse (pulse 2–10). All values are means ± SEM. Statistics with Student t test: *P < 0.05; **P < 0.01; ***P < 0.001.

Surprisingly, both the fast and slow phases of insulin release from female (WT or α₂δ-1^−/−) islets were significantly higher than insulin release from male islets of the same genotype. As a consequence, α₂δ-1^−/− female islets (Fig. 7E–G, red) secrete more than twice as much insulin as α₂δ-1^−/− male islets in response to glucose (integral of insulin release: WT females 135.8 ± 20.2, males 60.1 ± 10.2, P = 0.007; α₂δ-1^−/− females 34.25 ± 7.3, males 14.5 ± 2.7, P = 0.04).

Discussion

A broad spectrum of symptoms and tested parameters demonstrate that loss of α₂δ-1 Ca²⁺ channel subunit increases the susceptibility to diabetes in mice. Consistent with the role of α₂δ-1 in functional membrane expression of Ca²⁺ channel shown in other cell types (35–37), our data show that Ca²⁺ influx is equally reduced in β-cells of both sexes. In KO males, this results in reduced insulin release and in an escalation of diabetes with severe β-cell loss. Whereas a more robust insulin release in female mice enables most KO females to cope with the reduced Ca²⁺ signal, and to sufficiently regulate blood glucose levels and maintain β-cell mass. The finding that an indiscriminate reduction of Ca²⁺ currents of all HVCC isoforms of β-cells leads to a complex sex-specific diabetes phenotype is novel. Similar to the situation in human diabetes, genetic variations affecting the basal efficiency of insulin release (38) contributes to the manifestation of diabetes in the α₂δ-1^−/− mice. This could also explain why an overt diabetic phenotype has not been noted up to 12 weeks of age, in an α₂δ-1^−/− mouse colony on a C57BL/6J genetic background (28) (A.C. Dolphin, personal communication).

As demonstrated here for mice, also in human pancreatic islets, α₂δ-1 is the dominant α₂δ subunit isoform (22,39,40). Although several genomic aberrations of the CACNA2D1 gene have been associated with epilepsy and intellectual disability (23), short QT syndromes (24), and cardiac arrhythmias (26), the role of α₂δ subunits in diabetes has hitherto not been analyzed in these patients. Our current findings suggest that patients carrying CACNA2D1 loss-of-function mutations, in particular males, should be closely monitored for impaired glucose tolerance.

Primary Cause of Diabetes in α₂δ-1^−/− Mice Is β-Cell Dysfunction

Several lines of evidence indicate that the primary cause of diabetes in α₂δ-1^−/− mice is reduced insulin secretion. Ca²⁺ currents in α₂δ-1^−/− β-cells and Ca²⁺ influx–induced exocytosis are reduced. The first phase of the glucose-induced [Ca²⁺]i signal in α₂δ-1^−/− islets is delayed and the glucose-induced insulin release is reduced not only in affected α₂δ-1^−/− males but also in unaffected females and in young mice prior to the development of severe clinical signs. In theory, the general KO of the α₂δ-1 subunit could also affect glucose regulation by islet-independent mechanisms. In particular, in skeletal muscle, where α₂δ-1 is the exclusive α₂δ-1 isoform, its KO might impair the regulation of glucose uptake. However, in α₂δ-1^−/− mice, insulin sensitivity was not different from normal controls.

Another cause for diabetes in α₂δ-1^−/− mice could be reduced β-cell mass due to impaired β-cell neogenesis or survival. However, our data show that in young males and older asymptomatic females, the number of pancreatic islets is not affected (Fig. 3A–D), indicating that impaired generation or postnatal survival of β-cells cannot be the primary cause of the reduced insulin release. The dramatic loss of β-cells in older males is most probably secondary to impaired insulin release and blood glucose regulation (41). Furthermore, if the α₂δ-1 subunit is required for pancreatic endocrine cell survival, then an equal reduction of a-cells would be expected in α₂δ-1^−/− males, which was not observed.

α₂δ-1 Knockout Causes Reduced Ca²⁺ Currents Through All HVCC Isoforms

Our observation that a substantial reduction of Ca²⁺ currents occurred in β-cells of young α₂δ-1^−/− mice and that a similar current reduction was seen in asymptomatic females excludes the possibility that this effect on the Ca²⁺ currents was the result, rather than the cause, of diabetes (42). In pancreatic islets of the α₂δ-1^−/− mice, the upregulation of α₂δ-2 and -3 transcripts is way too small to quantitatively compensate for the loss of α₂δ-1. Our data also suggest that in normal β-cells, all HVCC types associate with the α₂δ-1 isoform since the α₂δ-1 KO equally reduces Ca²⁺ currents carried by all examined HVCC types. Moreover, the requirement of α₂δ-1 for the membrane expression of multiple HVCCs also explains the complex effects of α₂δ-1 KO on the different phases of Ca²⁺ dynamics and insulin secretion.

α₂δ-1 Deletion Alters Ca²⁺ Dynamics and Reduces First- and Second-Phase Insulin Secretion

Ca_V1.2 channel activation is required for insulin vesicle fusion (3) and critically contributes to the initiation of electrical activity in β-cells (2), whereas the activation of Ca_V2.3 channels plays a role in maintaining the Ca²⁺ signaling and insulin release in the slow-sustained second phase (9,10). Consistent with the importance of α₂δ-1 for the function of both of these HVCCs, the Ca²⁺ dynamics of the first phase was significantly retarded and both phases of insulin release were strongly reduced in islets of KO males and females. Strikingly, in male α₂δ-1^−/− islets, the reduction of Ca²⁺ dynamics and insulin release was greater in magnitude compared with what has earlier been demonstrated for male Ca_V1.2^−/− or Ca_V2.3^−/− mouse models alone (2,9). As α₂δ-1 appears to be the common constituent of multiple HVCCs in β-cells, its loss has more severe consequences than that of an individual HVCC α₁-subunit, and consequently genetic defects of α₂δ-1 in patients may have more detrimental effects on blood glucose regulation than the loss of function of a single pore-forming HVCC subunit.

Although both phases of insulin release and the first-phase Ca²⁺ signal were equally affected in KO males and females in response to a glucose challenge, the strong sex difference in the Ca²⁺ dynamics and insulin release in the WT animals explains the lower incidence of diabetes in KO females. First, the second-phase Ca²⁺ signal was not different in α₂δ-1^−/− and WT females. Notably, this type of measurement does not reflect Ca²⁺ dynamics of single β-cells but the complex spatio-temporal spread of the Ca²⁺ signal through the entire islet (43,44). Second, female islets secrete more insulin than male islets, and even though the α₂δ-1 deletion strongly reduced insulin release in both sexes, the remaining release in female KO is similar to that in WT males and appears to be sufficient to regulate blood glucose levels and prevent β-cell loss and the escalation of the diabetic phenotype. This interpretation is consistent with the higher efficiency of glucose utilization in the IPGTT test in KO females compared with KO males. Our data also support the notion that the observed sex difference in the incidence of diabetes is caused by differences in β-cell function and is not a sex-specific metabolic phenomenon independent of the islet β-cell. The insulin sensitivity was not sex specific, and the differences in insulin release were observed in isolated islets therefore in the absence of any acute sex hormone regulation.

Conclusion

The current study demonstrates a key role of the α₂δ-1 HVCC subunit in insulin metabolism–secretion coupling. The α₂δ-1 subunit is required for proper β-cell Ca²⁺ influx through multiple HVCC isoforms, and its absence results in reduced insulin secretion and impaired glucose tolerance. Our findings highlight that the control of HVCC currents by α₂δ-1 influences the in vivo insulin secretion and may have direct clinical relevance for patients carrying loss-of-function mutations in the CACNA2D1 gene (23,24,26).