Advertisement

Springer Nature is making SARS-CoV-2 and COVID-19 research free. View research | View latest news | Sign up for updates

The human L-type calcium channel Cav1.3 regulates insulin release and polymorphisms in CACNA1D associate with type 2 diabetes

Abstract

Aims/hypothesis

Voltage-gated calcium channels of the L-type have been shown to be essential for rodent pancreatic beta cell function, but data about their presence and regulation in humans are incomplete. We therefore sought to elucidate which L-type channel isoform is functionally important and its association with inherited diabetes-related phenotypes.

Methods

Beta cells of human islets from cadaver donors were enriched using FACS to study the expression of the genes encoding voltage-gated calcium channel (Cav)1.2 and Cav1.3 by absolute quantitative PCR in whole human and rat islets, as well as in clonal cells. Single-cell exocytosis was monitored as increases in cell capacitance after treatment with small interfering (si)RNA against CACNA1D (which encodes Cav1.3). Three single nucleotide polymorphisms (SNPs) were genotyped in 8,987 non-diabetic and 2,830 type 2 diabetic individuals from Finland and Sweden and analysed for associations with type 2 diabetes and insulin phenotypes.

Results

In FACS-enriched human beta cells, CACNA1D mRNA expression exceeded that of CACNA1C (which encodes Cav1.2) by approximately 60-fold and was decreased in islets from type 2 diabetes patients. The latter coincided with diminished secretion of insulin in vitro. CACNA1D siRNA reduced glucose-stimulated insulin release in INS-1 832/13 cells and exocytosis in human beta cells. Phenotype/genotype associations of three SNPs in the CACNA1D gene revealed an association between the C allele of the SNP rs312480 and reduced mRNA expression, as well as decreased insulin secretion in vivo, whereas both rs312486/G and rs9841978/G were associated with type 2 diabetes.

Conclusion/interpretation

We conclude that the L-type calcium channel Cav1.3 is important in human glucose-induced insulin secretion, and common variants in CACNA1D might contribute to type 2 diabetes.

Introduction

Insulin secretion requires the presence and activity of voltage-gated calcium channels (VGCCs [1]). Various VGCC subclasses are described and the L-type class (L-VGCC) has been considered essential for insulin secretion [2]. However, which of the two main L-type isoforms (Cav1.2 and Cav1.3) is the more important appears to be species dependent and to differ between rodents and humans [3]. Conclusive data showing gene expression in human beta cells has not previously been available.

In mice, global knockout of Cacna1c (which encodes Cav1.2) is lethal [4], whereas beta cell-specific disruption has demonstrated that Cav1.2 is critical for insulin release [5, 6], especially during the first phase [7]. Cav1.3, on the other hand, has been proposed to play a central role during murine postnatal beta cell generation and proliferation [8], but is also differentially regulated in a diabetogenic diet mouse model [9]. This isoform is characterised by a more negative activation threshold [10] and decreased calcium-channel blocker sensitivity [11] compared with Cav1.2, and has been suggested to be the main L-VGCC responsible for insulin release in rat beta cells [12, 13].

In humans, perturbed L-VGCC expression and function can cause disease, as exemplified by Timothy syndrome (caused by alterations in Cav1.2, encoded by CACNA1C) and congenital stationary night blindness (caused by dysfunctional Cav1.4, encoded by CACNA1F [14]). For human islets, mRNA expression of CACNA1D (encoding Cav1.3 protein) has been reported to exceed that of CACNA1C by tenfold [15]. Human genetic studies also point to CACNA1D being the most significant isoform for type 2 diabetes mellitus. A comparison between DNA from healthy individuals and those with type 2 diabetes revealed that, in the latter, CACNA1D carried an additional methionine-encoding triplet immediately downstream of the 5′ untranslated region [16, 17]. Although the functional consequences were not clarified, this finding supports the hypothesis of a causal link between genetic variations in the Cav1.3-encoding gene CACNA1D and risk for type 2 diabetes.

To evaluate this idea, we analysed data from genome-wide association studies (GWAS) for type 2 diabetes and quantitative traits [18, 19] in which we detected associations of single nucleotide polymorphisms (SNPs) in CACNA1D with type 2 diabetes. If and how these SNPs translate into changes in beta cell function have until now remained unanswered. Using genetics, molecular biology and electrophysiological approaches, we here present the first study determining the functional role of Cav1.3 for insulin secretion in the human beta cell and for human type 2 diabetes.

Methods

Real-time PCR

RNA was extracted and reverse transcribed as previously described [20]. For absolute quantification, we produced standard curves using custom-designed DNA oligonucleotides (4 nM; idtDNA, Coralville, IA, USA). Transcript numbers are expressed as transcripts/ng reverse transcribed mRNA. For primer sequences, see electronic supplementary material (ESM) Table 1.

FACS of beta cells

Beta cells from dispersed human islets were enriched using the Zn2+ dye Newport Green [21]. Apoptotic cells were accounted for (7-AAD staining) and islet cells were sorted using BD Biosciences FACSAria (BD Biosciences, San Jose, CA, USA). Beta cell enrichment was verified by immunohistochemistry and insulin-positive cells accounted for 90 ± 5% of all cells in the beta cell fractions, comparable with previous observations [22].

Microarrays of human pancreatic islet mRNA

RNA was isolated and prepared and the microarrays were performed and analysed as described in [23].

Only non-diabetic individuals were considered for correlations with CACNA1D expression. The genes that showed positive correlation (cut-off Pearson’s r ≥ 0.8) with CACNA1D expression (163 genes) were analysed using the DAVID database for functional annotation (david.abcc.ncifcrf.gov). The gene ontology category with the highest stringency for specific classification of gene product function (GOTERM_BP_5 in the gene ontology classification chart) was used and functions were sorted according to Fisher’s exact p value (minimum gene count threshold of 5). All microarray data are presented as arbitrary log2 units (AUs).

Immunocytochemistry and microscopy

Islets were treated as previously described [24] and incubated overnight at 4°C with anti-Cav1.3 (17 μg/ml, Cat. no. ACC-005, Lot no. AN-15; Alomone Labs, Jerusalem, Israel) and anti-insulin (1:200, Cat. no. 2263B65-1, Lot no. HS1761; EuroProxima, Arnhem, the Netherlands) or anti-glucagon (1:200, Cat. no. 2263B-GP310-1, Lot no. LN2051; EuroProxima) antibodies. Incubation with secondary antibodies anti-rabbit-DyLite488 and anti-guinea pig-Cy5 (1:200; Jackson ImmunoResearch, West Grove, PA, USA) for 4 h at RT and nuclei staining with Hoechst 33258 (1:1,000) was followed by visualisation in multitrack mode with a Zeiss 510 LSM confocal microscope with a C-Apochromat 40×/1.2 WDICIII lens (Carl Zeiss, Oberkochen, Germany).

Human study populations

Non-diabetic participants from the Botnia Family Study (IVGTT, n = 766) and the population-based Prevalence, Prediction and Prevention of Diabetes (PPP)-Botnia Study (OGTT, n = 4,671) were used to investigate effects of genetic variants in CACNA1D on insulin secretion in vivo (Table 1) [25, 26]. First-phase insulin release was determined as previously described [25]. Insulin and glucose measurements were also performed as previously described [26]. Insulin readouts were converted (1 μU/ml = 6.945 pmol/l) and log e transformed. The Malmö case–control population contained 2,830 patients with diabetes from the Malmö Diabetes Registry [27] and 3,550 non-diabetic controls from the Malmö Diet and Cancer study (Table 1) [28]. All cases were of Scandinavian origin, with an age at diabetes onset of older than 35 years, C-peptide levels of 0.3 nmol/l or higher and no GAD antibodies. Controls had fasting blood glucose less than 5.5 mmol/l, HbA1c less than 6.0% (42 mmol/mol) and no known first-degree relatives with diabetes. The diagnosis of type 2 diabetes was based on WHO criteria [29]. All participants gave informed consent and the protocols were approved by the ethics committees of Jorbruksverket (Human islets and MCC), Central Hospital Vasa (Botnia) and Helsinki University Hospital (Botnia-PPP).

Table 1 Clinical characteristics of the study participants

SNP selection and genotyping

Genotyping was performed as previously described [23]. The SNPs analysed were not in linkage disequilibrium with each other and were chosen based on nominal associations with risk of type 2 diabetes in the Diabetes Genetics Initiative GWAS (rs9841978) [18] or the DIAGRAM + meta-analysis for type 2 diabetes (rs312486 and rs9841978) [19], or based on their location and functional implication (rs312480 [16, 17, 30]; rs312486 as proxy for rs312481 [31]) (see ESM Table 2 for details). The genotyping success rate was higher than 97% (5% random replicates) and Hardy–Weinberg equilibrium was met in all studied populations.

In vitro insulin-release measurements, cell and islet culture

Human pancreatic islets were supplied by the Nordic Network for Clinical Islet Transplantation (courtesy of O. Korsgren, Uppsala University, Uppsala, Sweden). Insulin-release measurements were conducted as previously described [32]. Culture of INS-1 832/13 cells and tests for glucose-induced insulin secretion (GSIS) were performed as previously described [20]. Diagnosis of type 2 diabetes was based on donor history records or as indicated otherwise (i.e. based on WHO criteria). HbA1c was determined from patient blood samples taken while on life support.

Days in culture had no influence on CACNA1D expression in human islets. Although islet purity correlated significantly with CACNA1D expression (p < 0.01), the mean purity between islets from control and diabetic participants was not significantly different (68 ± 2% in controls vs 64 ± 6% in diabetic participants).

Transfection with siRNA

The small interfering (si)RNA oligonucleotides (rat siRNA ID s132089; human s2286) were supplied by Ambion (Austin, TX, USA). The negative control siRNA was non-targeting siRNA#1 (Cat. no. D-001810-01-05) from Dharmacon (Lafayette, LA, USA), as was the transfection reagent Dharmacon-1 (Cat. no. T-2001). Transfection was performed as previously described [20], but a concentration of 25 μM was used in the current study.

Electrophysiology

Human islets were dispersed into single cells by incubation in Ca2+-free buffer followed by trituration and overnight incubation in RPMI1640 medium containing 5 mmol/l glucose. They were then transfected with siRNA (see above) with additional co-transfection of a green fluorescent oligonucleotide (Cat. no. 2013; Invitrogen, Carlsbad, CA, USA) to identify successfully transfected cells. After 48 h, cells were used for electrophysiological capacity measurements as previously described [15].

Ca2+ imaging

Imaging was performed using a Polychrome V monochromator (TILL Photonics, Graefeling, Germany) on a Nikon Eclipse Ti Microscope (Nikon, Tokyo, Japan). A ER-BOB-100 trigger, iXON3 camera and software iQ2 (Andor Technology, Belfast, UK) were used for control and detection. After 24 h culture in poly-L-lysine-coated glass-bottom dishes in indicated glucose concentrations, human islets were loaded with Fura-2 (2 μmol/l in KRB buffer containing 5 mmol/l glucose) for 35–40 min, left in the imaging system for 15 min while perfused with 37°C buffer prior to imaging. The exposure time was 50 ms at 2 s intervals. Islets were exposed to 5 mmol/l glucose for less than 5 min (time interval 1 [ti1]), to 20 mmol/l glucose for 10 min (ti2), 5 mmol/l glucose for 5 min (ti3), 5 mmol/l glucose and 70 mmol/l KCl for 5 min (ti4) and 5 mmol/l glucose for 5 min (ti5). The AUC for the above time intervals was calculated as \( \mathrm{AU}{{\mathrm{C}}_{\mathrm{ti}}}=\mathrm{S}\left[ {0.5\times \left( {{{\mathrm{t}}_{\mathrm{n}}}-{{\mathrm{t}}_{{\mathrm{n}-1}}}} \right)\times \left( {\mathrm{Ratio}340/38{0_{\mathrm{n}}}+\mathrm{Ratio}340/38{0_{{\mathrm{n}-1}}}} \right)} \right] \).

Statistical analyses

Data are presented as means ± SEM. Quantitative traits were calculated using linear regression analysis corrected for age, sex and BMI. Type 2 diabetes risk was assessed using logistic regression adjusted for age, sex and BMI and is expressed as OR. p values were adjusted for the number of SNPs tested. All other data were statistically evaluated using an unpaired two-tailed Student’s t test (unless otherwise indicated) after verification of equal variances and normal distribution of the data. For human islet quantitative (q)PCR and microarray data, an unpaired two-tailed t test assuming unequal variances was used with confirmation by Mann–Whitney U test. p < 0.05 was considered significant.

Results

CACNA1C and CACNA1D expression in islets and beta cells

Reports on L-VGCC gene expression are not conclusive and we therefore quantified expression of Cacna1c and Cacna1d in rat islets, INS-1 832/13 cells and CACNA1C and -D in human islets with qPCR. In whole rat islets, we found 998 ± 197 Cacna1c mRNA transcripts (given as transcripts per ng reverse-transcribed RNA) and 9,727.9 ± 2,629.8 Cacna1d transcripts (Fig. 1a). Expression in INS-1 832/13 cells was similar to that observed in rat islets, with 1,580 ± 337 transcripts for Cacna1c and 4,266 ± 749 copies of Cacna1d (Fig. 1b). We then enriched beta cells from three islet donors by FACS using dispersed human islet cells and the Zn2+ dye Newport Green DCF [21]. For comparison, transcripts were also quantified in undispersed human islets of the same individuals, and for CACNA1C they amounted to 7 ± 2 (Fig. 1c) compared with 19 ± 1 in beta cells alone (Fig. 1d). The number of CACNA1D mRNA molecules amounted to 550 ± 30 in islets (Fig. 1c) and 1,130 ± 23 in human beta cells (Fig. 1d), suggesting CACNA1D transcripts to be mainly localised to beta cells. Staining for Cav1.3 protein in whole islets revealed Cav1.3 immunoreactivity not only in insulin-positive (beta) cells (Fig. 1i, k), but also in alpha cells (Fig. 1j).

Fig. 1
figure1

Expression of Cacna1c/CACNA1C (encoding rat/human Cav1.2) and Cacna1d/CACNA1D (rat/human Cav1.3). (a, b) mRNA expression of Cacna1d (grey bar) exceeds that of Cacna1c (white bar) in rat islets (a) as well as in INS-1 832/13 cells (b). The same relationship, but with greater difference, is found in human islets (c) and in human beta cells (d), where CACNA1D (grey bar) dominates over CACNA1C (white bar). (e, f) Microarray studies revealed that expression of CACNA1D (e) is reduced in islets of individuals with type 2 diabetes (grey bars) compared with controls (white bars), confirmed by qPCR (f). (g) Expression of CACNA1D is also reduced in islets from human organ donors with high HbA1c (≥6.5% [48 mmol/mol]; n = 6) as compared with those with low HbA1c (<6.5% [48 mmol/mol]; n = 42). (h) Expression of CACNA1C is not significantly changed but directionality is the same as for CACNA1D. (i,j) Cav1.3 immunoreactivity (green) is detected in insulin-positive cells (red, i) and in glucagon-positive cells of control islets (red, j). Note that only a few glucagon-positive cells are detected. (i) Inset: control with omitted primary antibody and final Hoechst 33258 staining. Scale bars 50 μm. (k) A single cell showing distinct staining for insulin (red) and Cav1.3 (green). Scale bar 5 μm. (l) Some small islets are Cav1.3 positive but lack insulin protein. Scale bar 50 μm. All data are means ± SEM. *p < 0.05. AU, arbitrary units

We then asked if CACNA1D expression was different in the islets of type 2 diabetes patients compared with controls. In microarrays CACNA1D mRNA was reduced in type 2 diabetes islets compared with controls (7.27 ± 0.16 AU in 10 cases vs 7.69 ± 0.55 AU in 66 controls; p = 0.03; t test; Fig. 1e). CACNA1C expression was not significantly changed but directionality was the same (6.22 ± 1.39 AU in cases vs 6.47 ± 0.04 in controls; Fig. 1h). Interestingly, the expression of CACNA1C and CACNA1D showed strong covariation (Pearson’s r = 0.9; p = 2.5 × 10−28; not shown). The downregulation of CACNA1D in type 2 diabetes islets was confirmed with absolute qPCR (1,235 ± 64 CACNA1D transcripts in 63 controls vs 754 ± 93 in seven cases; p = 2 × 10−4, t test; Fig. 1f). Similar results were obtained using HbA1c levels to define diabetic status, with CACNA1D expression being 50% lower in individuals with HbA1c of 6.5% (48 mmol/mol) or higher (626 ± 133 vs 1,250 ± 88 transcripts when HbA1c < 6.5% [48 mmol/mol]; p = 0.008, t test; Fig. 1g). We found no significant change in INS expression in type 2 diabetes islets as compared with controls (2−ΔCt INS relative to beta-actin 3.4 × 10−5 ± 1.5 × 10−5 in cases vs 4.6 × 10−5 ± 6.7 × 10−6 in controls, p = 0.25). This suggests that the reduced CACNA1D expression in islets from hyperglycaemic/type 2 diabetes donors cannot be attributed to reduced islet mass or reduced beta cell fraction within the islets, in agreement with previous thorough investigations of the same islet donor cohort [33, 34].

Effect of genetic variants in CACNA1D on insulin secretion and type 2 diabetes

The expression of the related L-VGCC gene CACNA1C has been shown to be genetically determined by SNPs [35], giving a functional explanation for associations of CACNA1C genotypes with neurological disorders [36, 37]. Similarly, to determine if the changes in expression of the Cav1.3 Ca2+ channel gene CACNA1D that we observed in human pancreatic islets are also influenced by genetic variations, we studied SNPs in CACNA1D and their associations with gene expression as well as with insulin release and type 2 diabetes.

We used literature searches and previously published GWAS studies to select the best candidate SNPs in CACNA1D ±50 kb (see Methods and ESM Table 2 for details).

The first marker analysed was rs312480 and the C allele associated with decreased 30 min insulin concentration (IVGTT; β = −0.103; p = 0.036, Table 2 and Fig. 2a). The same directionality (β) was observed for 30 min insulin during the Botnia-OGTT Study (Table 2), although values were not statistically significant. This suggests that the additional mechanisms of insulin release that are in operation when administering glucose via the oral route (e.g. the incretin effect) may obscure the direct effects of rs312480 on 30 min insulin values observed during IVGTT. Although we found no significant effect on fasting insulin in the smaller Botnia-OGTT Study, we detected decreased fasting insulin values in the OGTT of the larger PPP-Botnia Study (β = −0.047; p = 0.048) due to the larger sample size and therefore increased power.

Table 2 Effects of CACNA1D SNP rs312480 on insulin secretion during IVGTT and OGTT analysed by linear regression
Fig. 2
figure2

(a) rs312480 has effects on in vivo and in vitro insulin levels and on mRNA expression. Thirty minute insulin levels (standardised residual corrected for age, sex and BMI shown) are reduced in rs312480/C allele carriers of the Botnia IVGTT Study (a) as well as in human islet batch incubations (b). (c) CACNA1D expression is decreased in the same individuals as in (b). All data are means ± SEM. *p < 0.05. AU, arbitrary units

In line with these in vivo findings, rs312480/C showed reduced in vitro insulin release in batch incubations at high glucose (16.7 mmol/l; CC 1.4 ± 0.1 vs CT 2.2 ± 0.3 ng/islet × h, p = 0.04; T/T n.a.; Fig. 2b). Microarray data from islets of the same individuals revealed a concomitant reduction in CACNA1D mRNA in the carriers of CC genotypes (CC 7.6 ± 0.1 vs CT 8.0 ± 0.2 AU, p = 0.03; Fig. 2c).

Neither the variant rs9841978 nor rs312486 showed a significant association with insulin secretion (ESM Table 3) or with changes in gene expression. However, we next wanted to examine if any of these SNPs in CACNA1D that were previously suggested to be associated with disease would also have an effect on risk of type 2 diabetes. In the Malmö case–control population, including 6,570 individuals (2,830 type 2 diabetes cases and 3,740 controls), rs9841978/G (OR 1.16, 95% CI 1.06, 1.27; p = 0.003) and rs312486/G (OR 1.17, 95% CI 1.06, 1.3; p = 1.9 × 10−4) were indeed associated with an increased risk of type 2 diabetes. The genotype frequencies for control (type 2 diabetes) individuals for rs312486 were 61.9% (66.9%) for G/G, 33.7% (29.4%) for G/C and 4.4% (3.7%) for C/C. For rs9841978, they amounted to 48.1% (53.4%) for G/G, 41.9% (38.2%) for G/A and 10.1% (8.4%) for A/A genotype carriers. We verified our findings by meta-analysis of DIAGRAM + (see ESM Table 2) and our own results. For rs312486/G, we found an OR of 1.11 (95% CI 1.06, 1.16 p = 1.1 × 10−5) and for rs9841978/G the OR was 1.08 (95% CI 1.04, 1.12; p = 1.1 × 10−4). The reason why the markers rs9841978 and rs312486 exhibited no significant association with insulin secretion or gene expression, but an increased risk of type 2 diabetes, remains unclear, but may be because the risk of type 2 diabetes is not solely related to impaired beta cell function or because of confounding effects on survival.

Glucose effects

Given that CACNA1D transcript numbers were decreased in islets of individuals with type 2 diabetes, we next investigated the possible direct effects of glucose on CACNA1D expression in human islets. After 24 h incubation, insulin concentrations in the medium were 5.5-fold higher at 20 mmol/l glucose as compared with 5 mmol/l glucose, but in the presence of the L-type calcium-channel blocker isradipine the islets failed to release insulin (p = 0.02, ESM Fig. 1), as also reported elsewhere [15].

After 24 h islets were rested at 5 mmol/l glucose for 1 h then stimulated with 20 mmol/l glucose. Islets that had been cultured for 24 h in 20 mmol/l glucose now released significantly more insulin than islets cultured at basal (5 mmol/l) glucose (4.2 ± 0.4-fold vs 2.6 ± 0.4-fold, p < 0.01; Fig. 3a). The presence of isradipine in the 24 h incubation medium completely abolished GSIS in subsequent batch incubations. We next asked what effect this L-type channel inhibition had on CACNA1D mRNA expression. Absolute quantification revealed that high glucose increased CACNA1D transcript numbers approximately twofold, whereas expression levels remained unaffected in the presence of isradipine (Fig. 3b). Increased CACNA1D expression coincided with elevated basal [Ca2+]i levels (2.14 ± 0.04 AUCi for islets cultured for 24 h at 20 mmol/l vs 1.68 ± 0.04 when tested at 5 mmol/l glucose; p = 8 × 10−4; Fig. 3c, d), an exaggerated response to 20 mmol/l glucose (6.09 ± 0.09 for 20 mmol/l glucose vs 4.5 ± 0.05 for 5 mmol/l; p = 1 × 10−5) and higher levels during (6.51 ± 0.39 vs 5.53 ± 0.16; p = 0.046) and after 70 mmol/l KCl stimulation (1.84 ± 0.05 vs 1.44 ± 0.05; p = 2 × 10−3). Typical [Ca2+]i oscillations during high-glucose stimulation were absent in the islets that had been incubated at 20 mmol/l glucose for 24 h (Fig. 3d). The elevated expression of CACNA1D by long-term 24 h glucose treatment is thus functionally important and clearly affects [Ca2+]i homeostasis.

Fig. 3
figure3

CACNA1D expression changes upon incubation with 20 mmol/l glucose (Gluc.). Incubation of control human islets in medium with 20 mmol/l glucose for 24 h resulted in an increased fold insulin release (a, light grey bars) in a batch incubation and increased CACNA1D expression (b, light grey bars), compared with the 5 mmol/l glucose condition (white bars), unless the L-type Ca2+ channel inhibitor isradipine (Isr.) had been present for 24 h (a, b; dark grey bars). No isradipine was added during the batch incubation. In parallel, intracellular Ca2+ levels were increased after 24 h in 20 mmol/l glucose (c, grey bars, given as the sum of the AUC [A.U.C.ti] in the given time intervals [5 or 10 min]) compared with 24 h in 5 mmol/l glucose (white bars). The example shown in (d) with inset depicts the absence of [Ca2+]i oscillations during high-glucose stimulation in the islets incubated for 24 h at 20 mmol/l glucose (black traces) compared with 24 h in 5 mmol/l glucose (grey traces). (e) siRNA against Cacna1d reduced Cacna1d mRNA expression (grey bars) but not Cacna1c expression (white bars). (f) Glucose-induced insulin release (16.7 mmol/l) was reduced in siCav1.3-treated INS-1 832/13 cells (grey bars), whereas basal release was unaffected (2.8 mmol/l; white bars). (g) In analogy, in human beta cells, exocytosis was reduced after siCav1.3 treatment (white bar) compared with siContr (grey bar). A typical example is shown in (h), with the stimulation protocol being illustrated. (i) Cav1.3 knockdown (siCav1.3) in INS-1 832/13 cells also reduced Cav1.3 protein (Cav1.3) without an effect on insulin protein expression (Insulin) when compared with controls (siContr). Absence of Cav1.3 immunoreactivity upon pre-incubation with Cav1.3 peptide (i, siContr + peptide) verifies the specificity of the antibody. The intensity profiles depict the intensity of the red lines drawn in the corresponding left panels (Cav1.3). All data are means ± SEM. *p < 0.05 (Student’s t test); siContr, control siRNA

As we detected decreased CACNA1D expression in the islets of individuals with type 2 diabetes, we wanted to examine whether the presence of CACNA1D transcripts was a prerequisite for appropriate glucose-induced insulin release. We used siRNAs against human CACNA1D (siCACNA1D), rat Cacna1d (siCacna1d) and the rat beta cell line INS-1 832/13, a model for the study of GSIS [38]. The two siRNAs tested (designed for human and rat CACNA1D, but both targeting rat Cacna1d) were comparably efficient and after 48 h Cacna1d expression in INS-1 832/13 cells was decreased by 72 ± 6% (Fig. 3e, shown only for siRNA against rat Cacna1d; p = 0.02) compared with the non-targeting siRNA. Cav1.3 protein was also decreased (Fig. 3i). We then tested for effects on GSIS. While basal release was not affected (Fig. 3f; 1.5 ± 0.1 ng insulin/mg protein × h in siCacna1d-treated vs 1.8 ± 0.3 ng insulin/mg protein × h in control cells), the response to high glucose was significantly reduced in cells treated with siCacna1d (Fig. 3f; 3.8 ± 0.4 ng/mg protein × h in siCacna1d vs 6.7 ± 0.7 ng/mg protein × h in control; p = 0.0026). Human beta cells responded in a similar manner. After 48 h incubation with siCACNA1D, depolarisation-evoked single-cell exocytosis was significantly reduced when compared with control cells (Fig. 3g, h; 41 ± 4 fF in siCACNA1D vs 68 ± 8 fF in control cells; p = 0.009). The difference was significant for each of the ten depolarisations (not shown). Exocytosis was reduced by about 40%, comparable with the reduction of GSIS in the rat beta cell line.

CACNA1D and the exocytotic microdomain

Our data indicate that the appropriate expression of CACNA1D and function of Cav1.3 are important determinants of GSIS. To further examine the role of Cav1.3 in the exocytotic process, we extended our microarray studies by exploring the genes that positively correlate with CACNA1D expression (cut-off Pearson’s r ≥0.8, 163 genes). We found that genes coexpressed with CACNA1D were significantly enriched for particular cellular functions (Table 3) and the top-ranked category was exocytosis (p = 9 × 10−4). This is in agreement with the functional results and underlines the relevance of these findings.

Table 3 Gene ontology analysis of the genes positively correlating with CACNA1D expression in control human islets (Pearson’s r ≥0.8; 163 genes). The three highest scoring categories of gene product functions are depicted

Discussion

The L-type calcium channels of the beta cell have been extensively studied, but mostly in rodents. It therefore remains unclear which isoform is expressed and operational in the human beta cell. In this study, we set out to determine this, as well as possible associations between genetic variations and phenotypes related to type 2 diabetes. We demonstrate that Cav1.3 dominates over the related Cav1.2 channel on the mRNA level, and that Cav1.3 mRNA and protein are mainly localised to human beta cells. Second, we have shown that a SNP (rs312480) in CACNA1D, the gene encoding Cav1.3, can influence gene expression of the channel and affect insulin release, and that other SNPs (rs312486 and rs9841978) associate with type 2 diabetes. Furthermore, our data illustrate that the expression of CACNA1D mRNA is reduced in islets of patients with type 2 diabetes (Fig. 1) and that pharmacological or RNA interference-mediated inhibition of CACNA1D expression and/or Cav1.3 operation decreases GSIS and, on a single-cell level, exocytosis (Fig. 3).

Animal studies designed to investigate which L-type channel isoform is expressed and functional in beta cells have revealed that insulin release in the mouse mainly depends on Cav1.2 [6, 7, 39, 40], with a diminished first phase after ablation of the channel [7]. However, a compensatory upregulation of Cav1.2 upon Cacna1d knockout, as detected by another group [8], could not be ruled out [39]. Studies in rats claimed Cav1.3 to be the major L-VGCC involved [12], which is similar to what is proposed for humans [15], although neither study quantified protein levels. We detected Cav1.3 transcripts and protein not only in whole islets, but also found both to be enriched in beta cells, although some protein was also detected in alpha cells. We also identified a common variant in CACNA1D (SNP rs312480) to be associated with reduced fasting insulin and putatively also with 30 min insulin levels, indicating that this isoform might not strictly participate in phasic but also in basal insulin secretion. In fact, Cav1.3, with its lower activation threshold compared with Cav1.2 [10], has been suggested to be involved in pacemaking in the heart [41] and in the neuroendocrine chromaffin cell [42]. Considering that the neuroendocrine beta cell also fires action potentials in the physiological resting (or preprandial) condition [43], Cav1.3, with its low activation threshold, might be a good candidate for these events. In the glutamate-releasing inner-ear hair cell, Cav1.3 has been shown to be responsible for the generation of spontaneous action potentials, important for the maturation of synaptic connections within the developing cochlea [44]. In a remarkable analogy, the postnatal maturation of mouse beta cells has also been suggested to be dependent on the presence of Cav1.3 [8], although spontaneous spiking during this process has never been investigated. Reflecting on our finding that a polymorphism in the gene encoding for Cav1.3 leads to decreased expression and reduced insulin release, our present data suggest that this is not related to a decrease in islet mass or beta cell proportion. The possibility that the genotype is associated with the capacity for postnatal human beta cell expansion, for instance in insulin resistance, requires live determination and longitudinal studies of beta cell mass, techniques that remain to be developed for application in humans [45]. Nevertheless, we here demonstrate that in beta cells of adult humans, Cav1.3 is important for insulin exocytosis as knockdown of its expression also reduced insulin vesicle fusion events (Fig. 3f, g). This is in accordance with findings from Braun and colleagues [15], where the L-type blocker isradipine abolished glucose-induced insulin release in human islet batch incubations and reduced exocytosis in patch experiments, although the latter was even more suppressed by P/Q-type channel blockers. Nevertheless, isradipine completely inhibited glucose-induced action potentials, suggesting that a single depolarisation in voltage-clamp mode strongly underestimates the pharmacological effect of L-type channel blockage on beta cell function.

We also found that Cav1.3 knockdown reduced insulin release in INS-1 832/13 cells. A previous study suggested that Cav1.2 knockdown diminished GSIS [46] but as we did not test siRNA against Cav1.2, it may well be that both channels are crucial for appropriate insulin release in these cells.

Our study also reveals that human CACNA1D expression responds to a change in glucose levels, a phenomenon that has previously been observed in rats [13]. No data on the effects of glucose on Cacna1d expression in mice are available, but a several-week-long diabetogenic high-fat diet has been shown to decrease rather than increase Cacna1d expression [9]. However, we found increased levels after a high-glucose episode of just 24 h, whereas after 48 h Iwashima and colleagues detected a decline (in rats) [13]. Although another species and time frame, this indicates a putative time- and activity-dependent regulation of CACNA1D expression. Indeed, for the related Cav1.2 channel, a mechanism has been put forward whereby calpain Ca2+ dependently cleaves an (among L-type VGCC well conserved [47]) autoinhibitory C-terminal domain of the channel. The cleaved C-terminus disinhibits the channel and acts as a transcription factor that controls expression of the Cav1.2 encoding gene CACNA1C [48]. Interestingly, when we pharmacologically inhibited Ca2+ influx through L-type channels and consequentially also insulin release by isradipine (Fig. 2a), CACNA1D expression was no longer augmented (Fig. 2b). This suggests suppression of the activity-dependent positive-feedback loop that otherwise would drive the rise in channel expression analogous to that proposed for Cacna1c (Cav1.2). We have indeed observed that calpain inhibitor-1 can reduce the level of CACNA1D mRNA in the high-glucose condition back to isradipine levels (not shown). The observed glucose-induced activity-dependent rise in CACNA1D expression correlated with increased basal Ca2+ levels, as well as higher levels under stimulation (Fig. 3c, d). Human islets cultured under high-glucose conditions for 48 h have previously been described to be desensitised due to increased basal Ca2+ levels and diminished slow [Ca2+]i oscillations, and thus to respond poorly to glucose stimulation [49]. Interestingly, after 24 h we also found diminished Ca2+ oscillations but hypersensitisation rather than desensitisation (Fig. 3d). It has been suggested that glucose causes hypersensitisation first, followed later by desensitisation and Ca2+-induced apoptosis [glucotoxicity; 50] and the two findings may thus be in line with each other.

It is tempting to speculate that pharmacological inhibition of excessive Ca2+ influx in prediabetic islets, in order to prevent Ca2+ induced hypersensitisation, may be an option for early-stage intervention that might at least delay the onset of type 2 diabetes.

Abbreviations

AU:

Arbitrary unit

Cav :

Voltage-gated calcium channel

GSIS:

Glucose-induced insulin secretion

GWAS:

Genome-wide association studies

L-VGCC:

L-type voltage-gated calcium channel

PPP:

Prevalence, Prediction and Prevention of Diabetes study

qPCR:

Quantitative PCR

siRNA:

Small interfering RNA

SNP:

Single nucleotide polymorphism

VGCC:

Voltage-gated calcium channel

References

  1. 1.

    Rorsman P (1997) The pancreatic beta-cell as a fuel sensor: an electrophysiologist’s viewpoint. Diabetologia 40:487–495

  2. 2.

    Davalli AM, Biancardi E, Pollo A et al (1996) Dihydropyridine-sensitive and -insensitive voltage-operated calcium channels participate in the control of glucose-induced insulin release from human pancreatic beta cells. J Endocrinol 150:195–203

  3. 3.

    Drews G, Krippeit-Drews P, Dufer M (2010) Electrophysiology of islet cells. Adv Exp Med Biol 654:115–163

  4. 4.

    Seisenberger C, Specht V, Welling A et al (2000) Functional embryonic cardiomyocytes after disruption of the L-type alpha1C (Cav1.2) calcium channel gene in the mouse. J Biol Chem 275:39193–39199

  5. 5.

    Barg S, Eliasson L, Renstrom E, Rorsman P (2002) A subset of 50 secretory granules in close contact with L-type Ca2+ channels accounts for first-phase insulin secretion in mouse beta-cells. Diabetes 51(Suppl 1):S74–S82

  6. 6.

    Vignali S, Leiss V, Karl R, Hofmann F, Welling A (2006) Characterization of voltage-dependent sodium and calcium channels in mouse pancreatic A- and B-cells. J Physiol 572:691–706

  7. 7.

    Schulla V, Renstrom E, Feil R et al (2003) Impaired insulin secretion and glucose tolerance in beta cell-selective Ca(v)1.2 Ca2+ channel null mice. EMBO J 22:3844–3854

  8. 8.

    Namkung Y, Skrypnyk N, Jeong MJ et al (2001) Requirement for the L-type Ca(2+) channel alpha(1D) subunit in postnatal pancreatic beta cell generation. J Clin Investig 108:1015–1022

  9. 9.

    Dreja T, Jovanovic Z, Rasche A et al (2009) Diet-induced gene expression of isolated pancreatic islets from a polygenic mouse model of the metabolic syndrome. Diabetologia 53:309–320

  10. 10.

    Koschak A, Reimer D, Huber I et al (2001) alpha 1D (Cav1.3) subunits can form l-type Ca2+ channels activating at negative voltages. J Biol Chem 276:22100–22106

  11. 11.

    Sinnegger-Brauns MJ, Huber IG, Koschak A et al (2009) Expression and 1,4-dihydropyridine-binding properties of brain L-type calcium channel isoforms. Mol Pharmacol 75:407–414

  12. 12.

    Seino S (1995) CACN4, the major alpha 1 subunit isoform of voltage-dependent calcium channels in pancreatic beta-cells: a minireview of current progress. Diabetes Res Clin Pract 28(Suppl):S99–S103

  13. 13.

    Iwashima Y, Pugh W, Depaoli AM et al (1993) Expression of calcium channel mRNAs in rat pancreatic islets and downregulation after glucose infusion. Diabetes 42:948–955

  14. 14.

    Striessnig J, Bolz HJ, Koschak A (2010) Channelopathies in Ca(v)1.1, Ca(v)1.3, and Ca(v)1.4 voltage-gated L-type Ca(2+) channels. Pflugers Arch 460:361–374

  15. 15.

    Braun M, Ramracheya R, Bengtsson M et al (2008) Voltage-gated ion channels in human pancreatic beta-cells: electrophysiological characterization and role in insulin secretion. Diabetes 57:1618–1628

  16. 16.

    Yamada Y, Masuda K, Li Q et al (1995) The structures of the human calcium channel alpha 1 subunit (CACNL1A2) and beta subunit (CACNLB3) genes. Genomics 27:312–319

  17. 17.

    Yamada Y, Kuroe A, Li Q et al (2001) Genomic variation in pancreatic ion channel genes in Japanese type 2 diabetic patients. Diabetes Metabol Res Rev 17:213–216

  18. 18.

    Saxena R, Voight BF, Lyssenko V et al (2007) Genome-wide association analysis identifies loci for type 2 diabetes and triglyceride levels. Science 316:1331–1336

  19. 19.

    Voight BF, Scott LJ, Steinthorsdottir V et al (2010) Twelve type 2 diabetes susceptibility loci identified through large-scale association analysis. Nat Genet 42:579–589

  20. 20.

    Reinbothe TM, Ivarsson R, Li DQ et al (2009) Glutaredoxin-1 mediates NADPH-dependent stimulation of calcium-dependent insulin secretion. Mol Endocrinol 23:893–900

  21. 21.

    Lukowiak B, Vandewalle B, Riachy R et al (2001) Identification and purification of functional human beta-cells by a new specific zinc-fluorescent probe. J Histochem Cytochem Offic J Histochem Soc 49:519–528

  22. 22.

    Kirkpatrick CL, Marchetti P, Purrello F et al (2010) Type 2 diabetes susceptibility gene expression in normal or diabetic sorted human alpha and beta cells: correlations with age or BMI of islet donors. PLoS One 5:e11053

  23. 23.

    Jonsson A, Isomaa B, Tuomi T et al (2009) A variant in the KCNQ1 gene predicts future type 2 diabetes and mediates impaired insulin secretion. Diabetes 58:2409–2413

  24. 24.

    Lyssenko V, Eliasson L, Kotova O et al (2011) Pleiotropic effects of GIP on islet function involve osteopontin. Diabetes 60:2424–2433

  25. 25.

    Groop L, Forsblom C, Lehtovirta M et al (1996) Metabolic consequences of a family history of NIDDM (the Botnia study): evidence for sex-specific parental effects. Diabetes 45:1585–1593

  26. 26.

    Pyykkönen A-J, Räikkönen K, Tuomi T, Eriksson JG, Groop L, Isomaa B (2010) Stressful life events and the metabolic syndrome. Diabetes Care 33:378–384

  27. 27.

    Lindholm E, Agardh E, Tuomi T, Groop L, Agardh CD (2001) Classifying diabetes according to the new WHO clinical stages. Eur J Epidemiol 17:983–989

  28. 28.

    Berglund G, Elmstahl S, Janzon L, Larsson SA (1993) The Malmo Diet and Cancer study. Design and feasibility. J Intern Med 233:45–51

  29. 29.

    World Health Organisation (2006) Definition and diagnosis of diabetes mellitus and intermediate hyperglycemia: report of a WHO/IDF Consultation. Available from http://www.who.int/entity/diabetes/publications/. Accessed 11 June 2011

  30. 30.

    Pickering BM, Willis AE (2005) The implications of structured 5′ untranslated regions on translation and disease. Semin Cell Dev Biol 16:39–47

  31. 31.

    Kamide K, Yang J, Matayoshi T et al (2009) Genetic polymorphisms of L-type calcium channel alpha1C and alpha1D subunit genes are associated with sensitivity to the antihypertensive effects of L-type dihydropyridine calcium-channel blockers. Circ J 73:732–740

  32. 32.

    Rosengren AH, Jokubka R, Tojjar D et al (2010) Overexpression of alpha2A-adrenergic receptors contributes to type 2 diabetes. Science 327:217–220

  33. 33.

    Taneera J, Jin Z, Jin Y, et al. (2012) γ-Aminobutyric acid (GABA) signalling in human pancreatic islets is altered in type 2 diabetes. Diabetologia 55:1985–1994

  34. 34.

    Taneera J, Lang S, Sharma A et al (2012) A systems genetics approach identifies genes and pathways for type 2 diabetes in human islets. Cell Metabol 16:122–134

  35. 35.

    Bigos KL, Mattay VS, Callicott JH et al (2010) Genetic variation in CACNA1C affects brain circuitries related to mental illness. Arch Gen Psychiatry 67:939–945

  36. 36.

    Sklar P, Smoller JW, Fan J et al (2008) Whole-genome association study of bipolar disorder. Mol Psychiatr 13:558–569

  37. 37.

    Green EK, Grozeva D, Jones I et al (2009) The bipolar disorder risk allele at CACNA1C also confers risk of recurrent major depression and of schizophrenia. Mol Psychiatr 15:1016–1022

  38. 38.

    Hohmeier HE, Mulder H, Chen G, Henkel-Rieger R, Prentki M, Newgard CB (2000) Isolation of INS-1-derived cell lines with robust ATP-sensitive K+ channel-dependent and -independent glucose-stimulated insulin secretion. Diabetes 49:424–430

  39. 39.

    Platzer J, Engel J, Schrott-Fischer A et al (2000) Congenital deafness and sinoatrial node dysfunction in mice lacking class D L-type Ca2+ channels. Cell 102:89–97

  40. 40.

    Sinnegger-Brauns MJ, Hetzenauer A, Huber IG et al (2004) Isoform-specific regulation of mood behavior and pancreatic beta cell and cardiovascular function by L-type Ca2+ channels. J Clin Investig 113:1430–1439

  41. 41.

    Mangoni ME, Nargeot J (2008) Genesis and regulation of the heart automaticity. Physiol Rev 88:919–982

  42. 42.

    Comunanza V, Marcantoni A, Vandael DH et al (2011) CaV1.3 as pacemaker channels in adrenal chromaffin cells: specific role on exo- and endocytosis? Channels 4:440–446

  43. 43.

    Misler S, Dickey A, Barnett DW (2005) Maintenance of stimulus-secretion coupling and single beta-cell function in cryopreserved-thawed human islets of Langerhans. Pflugers Arch 450:395–404

  44. 44.

    Brandt A, Striessnig J, Moser T (2003) CaV1.3 channels are essential for development and presynaptic activity of cochlear inner hair cells. J Neurosci 23:10832–10840

  45. 45.

    Ahlgren U, Gotthardt M (2010) Approaches for imaging islets: recent advances and future prospects. Adv Exp Med Biol 654:39–57

  46. 46.

    Nitert MD, Nagorny CL, Wendt A, Eliasson L, Mulder H (2008) CaV1.2 rather than CaV1.3 is coupled to glucose-stimulated insulin secretion in INS-1 832/13 cells. J Mol Endocrinol 41:1–11

  47. 47.

    Hulme JT, Konoki K, Lin TW et al (2005) Sites of proteolytic processing and noncovalent association of the distal C-terminal domain of CaV1.1 channels in skeletal muscle. Proc Natl Acad Sci U S A 102:5274–5279

  48. 48.

    Satin J, Schroder EA, Crump SM (2011) L-type calcium channel auto-regulation of transcription. Cell Calcium 49:306–313

  49. 49.

    Bjorklund A, Lansner A, Grill VE (2000) Glucose-induced [Ca2+]i abnormalities in human pancreatic islets: important role of overstimulation. Diabetes 49:1840–1848

  50. 50.

    Cnop M, Welsh N, Jonas JC, Jorns A, Lenzen S, Eizirik DL (2005) Mechanisms of pancreatic beta-cell death in type 1 and type 2 diabetes: many differences, few similarities. Diabetes 54(Suppl 2):S97–S107

Download references

Acknowledgements

We thank B.-M. S. Nilsson and A.-M. Veljanovska Ramsay (both Department of Clinical Sciences Malmö, Islet Pathophysiology, Lund University, Sweden) for their technical expertise and C. Ladenvall (Department of Clinical Sciences Malmö, Diabetes and Endocrinology, Lund University, Sweden) for advice on statistical analysis. INS-1 832/13 cells were kindly provided by H. Hohmeier (Department of Medicine, Duke University, Durham, NC, USA).

Funding

Imaging and qPCR equipment were financed by the Knut and Alice Wallenberg Foundation. T.R. was funded by the EU research training network Cavnet (MRTN-CT-2006-035367). E.R. is a Swedish Research Council senior researcher.

Contribution statement

TMR conducted research and wrote the manuscript. SA conducted research and contributed to revision of the manuscript. EA, TT and BI conducted research and reviewed the manuscript. VL and ER contributed to the concept and design of the study, and reviewed and edited the manuscript. All authors approved the final version of the manuscript.

Duality of interest

The authors declare that there is no duality of interest associated with this manuscript.

Author information

Correspondence to T. M. Reinbothe.

Electronic supplementary material

Below is the link to the electronic supplementary material.

ESM Fig. 1

PDF 1,349 kb

ESM Table 1

PDF 70 kb

ESM Table 2

PDF 88 kb

ESM Table 3

PDF 66 kb

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Reinbothe, T.M., Alkayyali, S., Ahlqvist, E. et al. The human L-type calcium channel Cav1.3 regulates insulin release and polymorphisms in CACNA1D associate with type 2 diabetes. Diabetologia 56, 340–349 (2013). https://doi.org/10.1007/s00125-012-2758-z

Download citation

Keywords

  • Beta cell
  • CACNA1D
  • Calcium
  • Cav1.3 channel
  • Diabetes
  • Exocytosis
  • Human
  • Insulin
  • Islets