Involvement of the Clock Gene Rev-erb alpha in the Regulation of Glucagon Secretion in Pancreatic Alpha-Cells

All animal work has been conducted according to national and international guidelines. All protocols were approved by the ethical committee of Miguel Hernandez University “Comisión de Ética en la Investigación Experimental” and specifically reviewed and approved this study (approval ID: IB-IQM-001-10).

The glucagon-releasing alphaTC1-9 cells were purchased from American Type Cultures Collection (ATCC CRL-2350; Barcelona, Spain). They were grown in DMEM (Invitrogen, Barcelona, Spain) supplemented with 4 mM L-glutamine, 16 mM glucose, 19 mM NaHCO3, 10% FBS, 15 mM HEPES and 0.1 mM non-essential amino acids. To measure circadian Rev-erb alpha gene expression in vitro, we performed a short treatment with 50% serum (serum shock) to the confluent, serum-starved alphaTC1-9 cells, according to previous studies [18], [19]. After 2 hours of serum shock, the medium was changed to DMEM serum-free medium and RNA was extracted every 6 hours during 48 hours for gene expression measurements. The experiments were done in serum-free medium to discriminate between oscillations in cell cycle and intrinsic circadian oscillations [18], [19]. Cell viability was performed in alphaTC1-9 cells in presence of 0.5 mM glucose and 11 mM glucose for 24 and 48 h using Trypan Blue stain 0.4% (Life technologies Ltd, Paisley, UK) and measured with Countless® automated cell counter (Invitrogen, Life technologies Ltd, Paisley, UK).

All protocols were approved by our animal Care Committee according to national regulations (ethical number: IB-IQM-001-10). Adult C57BL/6J mice were sacrificed at 8 weeks old by cervical dislocation and islets were then isolated by collagenase digestion [20] and hypothalamus and liver were collected for gene expression measurements. Mice were kept under 12 h:12 h light dark cycle (lights on at 07∶00 and lights off at 19∶00) with food ad libitum. For Ca²⁺ experiments, islets were dispersed into single cells in a Ca²⁺-deficient medium containing 0.5 mM EDTA and 0.05% trypsin followed by brief shaking. Then, cells were cultured overnight at 37°C in RPMI 1640 (Sigma, Madrid, Spain) supplemented with 10% fetal calf serum, 100 IU/ml penicillin, 0.1 mg/ml streptomycin and 11 mM D-glucose [21]. Either alphaTC1-9 cells or isolated mouse islet cells were loaded with Fura-2 (5 µM) for 1 hour at room temperature. Cells were placed on a perfusion chamber mounted on the microscope stage and perfused at a rate of 1.5 ml/min with a modified Ringer solution containing (in mM): 120 NaCl, 5 KCl, 25 NaHCO3, 1.1 MgCl2 and 2.5 CaCl2; pH = 7.4, gassed with 95% O2 and 5% CO2. Calcium signals were recorded using an inverted epifluorescence microscope (Zeiss, Axiovert 200) and a Hamamatsu Digital Camera C4742-95 (Hamamatsu Photonics, Barcelona, Spain). Fluorescence records were expressed as the ratio of fluorescence (R) at 340 nm and 380 nm (F340/F380). We analyzed the area-under-the-curve (AUC) of the whole period of stimulation using Origin software (Origin, Origin Lab Corporation, MA USA). This parameter is an indicator of the global calcium increase in the cells [22]. As previously reported, alpha-cells were identified by their spontaneous calcium signalling activity at low glucose concentrations and their characteristic response to epinephrine [13], [23], [24].

Cells were cultured for 2 days in DMEM prior to the measurements of glucagon secretion. Then, cells were incubated at 37°C in 0.5 ml Krebs–Ringer bicarbonate buffer supplemented with 15 mM HEPES, 0.5% BSA and 5.6 mM glucose, pH 7.4 for 30 min. Afterward, batches of 1×10⁶ alphaTC1-9 cells were incubated for 1 h at 37°C in 0.5 ml of Krebs–Ringer bicarbonate buffer supplemented with 0.5 mM, 11.2 mM glucose, the Rev-erb alpha agonist GSK4112 (Sigma St Louis, USA), Hemin (Sigma St Louis, USA), the Rev-erb alpha antagonist SR8278 (Sigma St Louis, USA) or the Nampt inhibitor FK866 (Cayman Chemical, Ann Arbor, USA). The supernatant was collected and glucagon secretion was measured. Cells were then lysed with 50 µl of lysis buffer (70% ethanol, 0.4% HCl at 30%, 24.6% distilled water) and incubated overnight at 4°C. Samples were centrifuged at 2500 rpm for 5 min and the supernatant was collected for glucagon content analysis. Glucagon was detected by ELISA using a commercial kit (YK090; Gentaur, Brussels, Belgium). Total protein was determined by the Bradford method.

Quantitative PCR assays were performed using CFX96 Real Time System (Biorad, Hercules, USA). Reactions were carried out in a final volume of 10 µl, containing 200 nM of each primer, 100 nM of endogenous control primer, 1 µl of cDNA and IQ™ Sybr® Green Supermix (Biorad, Hercules, USA). Samples were subjected to the following conditions: 10 min at 95°C, 40 cycles (10 s 95°C/7 s 60°C/12 s 72°C), and a melting curve of 63 to 95°C with a slope of 0.1°C/s. The housekeeping gene rplp0 (ribosomal protein large P0, alias 36B4) was used as the endogenous control for quantification [5]. The results were analyzed with CFX Manager Version 1.6 (Biorad, Hercules, USA) and values were expressed as the relative expression respect to control levels (2^−ΔΔct). Primers sequences are described in Table 1.

SiRNA treatment was performed in alphaTC1-9 cells as previously described [25], [26]. Cells were transfected overnight with 50 nM of siRNAs Silencer® Pre-designed Rev-erb alpha (Ambion, TX,USA) or 50 nM of Silencer® labelled negative control #2 siRNA (Ambion, TX, USA) in optiMEM® I (Invitrogen, Carlsbad, USA) culture medium without antibiotics and 1% of Lipofectamine 2000 (Invitrogen, Carlsbad, USA). The following Rev-erb alpha siRNA sequences were used (5′-3′): GCAUCGUUGUUCAACGUGAtt, sense; UCACGUUGAACAACGAUGCaa, antisense. After overnight incubation, the transfection medium was replaced by DMEM culture medium for 24 h before the start of the experiments.

Cell pellets were obtained by centrifuging at 1000×g for 10 min and resuspended in 50 µl of lysis buffer (Cell Signaling Technology, Danvers, MA). Cell extracts were subjected to sodium dodecyl sulfate (SDS) polyacrylamide gel electrophoresis (Mini-Protean®TGX™ Precast Gel, 4–20% gels, Biorad). Pre-stained SDS-PAGE standards were included for molecular mass estimation. The transfer to PVDF membranes was performed at 125 mA for 90 min in a buffer with 2.5 mM Tris base, 9 mM glycine, 20% methanol. After membranes were blocked with 2% non-fat dry milk, they were incubated with the following antibodies: rabbit polyclonal anti-actin (1∶1000; Sigma, Saint Louis, USA), anti-Rev-erb alpha (1∶500; Abcam, Cambridge, UK), pAMPK (Thr¹⁷²) (1∶1000; Abcam, Cambridge, UK) and Nampt antibody (1∶1000; Abcam, Cambridge, UK). Membranes were incubated with appropriate HRP-conjugated antibodies (Biorad, Hercules, USA). Protein bands were revealed by using the ECL western blot substrate (Thermo Fisher Scientific, Madrid, Spain). Intensity of the bands was quantified using Scion image software (Frederick, MD USA).

Isolated cells were placed in sorting buffer containing 1% BSA, 2.5 mM glucose, Ca²⁺, Mg²⁺ free PBS, 1 mM EDTA, 25 mM Hepes and transferred to a fluorescence activated cell sorter (FACS; BD FACSARIA SORP) equipped with 488 nm and 355 nm. The FACS was used to separate pancreatic beta-cells and alpha cells by cellular autofluorescence as recently reported by Kohler [27].

After sorting approximately 10.000 cells from each cell fraction were subjected to cytospin in order to perform immunofluorescence analysis. Cells are fixed in slides with Paraformaldehyde 4% and permeabilized with triton 0.2% for 45 minutes. Blocking was then performed with PBS-BSA 0.1% for 30 minutes at room temperature. Cells were incubated overnight at 4°C with insulin and glucagon antibodies (DAKO), and subsequently incubated with secondary antibodies Cy3 antiguineapig for insulin and Cy2 antirabbit for glucagon (Jackson Immunoresearch) and washed with PBS. Analysis performed with Image J program. Alpha-cell enriched fraction was composed of 80% alpha-cells and 4% beta-cells while the beta-cell enriched fraction had 75% beta-cells and no alpha-cells.

Data is shown as mean ± SEM. Student’s t test, one-way ANOVA or two-way ANOVA with Bonferroni correction were performed as appropriate with a level of significance p<0.05.

To analyze the mRNA expression of several clock genes, we used the mouse alpha-cell line alphaTC1-9 which has been validated as a good model to study alpha cell function [16], [25], [28]. The mRNA levels in this cell line were compared with those of the hypothalamus and liver. Rev-erb alpha expression in alphaTC1-9 cells was comparable with that in the hypothalamus whereas the expression levels in the liver were much higher (Fig. 1A). The expression of Clock, Bmal1 and Cry1 in alphaTC1-9 was similar to that in the hypothalamus and liver (S 1). Per-1 expression was lower in the liver compared with alphaTC1-9 cells whereas Per2 and Cry2 had lower expression levels in the hypothalamus (S 1). To study whether alpha-cells exhibit an oscillatory pattern of Rev-erb alpha expression along the day, we next performed mRNA measurements every 6 hours during a 48 h, as previously described [19]. To check that culture conditions were not affecting cell viability, we measured this parameter at 0.5 and 11 mM glucose after 24 and 48 hours of incubation. No differences were found in cell viability between both glucose concentrations in any of the time points measured (data not shown). Figure 1B shows that Rev-erb alpha mRNA levels oscillated in vitro at low glucose concentrations (0.5 mM) with an expression peak at ZT6 (Zeitgeber Time). Interestingly, in our experiments high glucose concentrations (11 mM) inhibited Rev-erb alpha expression levels. To further confirm the glucose effect on Rev-erb alpha at both the mRNA and protein levels, we performed similar experiments at ZT6. As expected, high glucose concentrations produced an inhibitory action on both Rev-erb alpha gene (Fig. 1C) and protein expression (Fig. 1D). We next studied the glucose effect at ZT6 in sorted mouse primary alpha-cells. In agreement with the results in alphaTC1-9 cells, high glucose concentrations down-regulated the Rev-erb alpha mRNA levels compared to the effect of 0.5 mM glucose in the alpha cell rich fraction of the primary isolated islet cells (Fig. 1E). These findings indicate that both alphaTC1-9 cells and mouse pancreatic alpha-cells express Rev-erb alpha and that glucose regulates the expression of this clock gene. Additionally, we show that alphaTC1-9 cells exhibit intrinsic circadian oscillations of Rev-erb alpha, indicating that alpha-cells of the endocrine pancreas have their own biological clock.

To study the functional role of Rev-erb alpha in alpha-cells, we used the siRNA technique to down-regulate this gene in mouse alphaTC1-9 cells. Gene silencing efficiency was about 60–70% compared with cells treated with a scramble siRNA (Sc; control siRNA) (Fig. 2A). At these concentrations of siRNA (50 nM) we previously observed no difference in cell viability between scramble and siRNA treated cells 13↗. The lower Rev-erb alpha expression was associated with a decrease in protein level (Fig. 2B). We next investigated whether Rev-erb alpha is involved in the regulation of glucagon secretion. Figure 2C shows the glucose effect on glucagon secretion from alphaTC1-9 cells after Rev-erb alpha silencing. As previously shown in isolated mouse islets and alphaTC1-9 cells [13], [16], glucagon secretion was stimulated at 0.5 mM glucose whereas secretion was inhibited at 11 mM glucose (Fig. 2C). After silencing the Rev-erb alpha gene in alphaTC1-9 cells, glucagon secretion at low glucose concentrations (0.5 mM) was decreased to the same extent as high glucose concentrations (11 mM) (Fig. 2C). No significant differences were observed either on glucagon content or glucagon gene expression after Rev-erb alpha silencing (data not shown). Interestingly, when we checked the mRNA expression of exocytotic genes such as Vamp3, Munch18, SNAP25 and Syntaxin1a, there was a down regulation of Munch18 and Syntaxin1a genes (and a tendency to decrease in VAMP3) in alphaTC1-9 cells treated with siRev compared with controls (Fig. 3A–D, respectively). Thus, the decreased glucagon release in cells with down-regulated Rev-erb alpha seems more related with changes in the mechanisms that allow the release of glucagon rather than a consequence of an altered glucagon synthesis.

The activity of Rev-erb alpha proteins can be modulated by its natural ligand heme [12], [29] or by the new synthetic Rev-erb alpha agonist GSK4112 and the antagonist SR8278 [30], [31]. To further evaluate the regulatory role of Rev-erb alpha in alpha-cell function, we checked the effect of GSK4112 and hemin on glucagon secretion in alphaTC1-9 cells. At 0.5 mM glucose the agonist GSK4112 had a stimulatory effect on glucagon secretion (Fig. 4A) but GSK4112 was not able to revert the inhibitory effects of 11 mM glucose. (Fig. 4A). On the other hand, the natural Rev-erbalpha agonist hemin stimulated glucagon secretion at both 0.5 mM and 11 mM glucose in isolated islets (S 2A). To test the biological effect of hemin, we measured Alas-1 (δ-aminolevulinate synthase 1) a rate limiting enzyme in the heme biosynthetic pathway. As expected the mRNA levels of Alas-1 was decreased in presence of hemin in alphaTC1-9 cells (S 2B).

The stimulus-secretion coupling of pancreatic alpha-cells is still controversial [20], [23], [32], [33], [34], [35], [36]. However, it is well accepted that glucagon secretion is a calcium-dependent mechanism and changes in calcium signaling regulate the exocytotic process in alpha-cells [34], [35], [37], [38]. Therefore, we checked whether the stimulatory effect of the Rev-erb alpha agonist GSK4112 on glucagon secretion could be via modulation of intracellular calcium levels. At low glucose levels (0.5 mM) alphaTC1-9 cells exhibited intracellular calcium oscillations (Fig. 4B). Addition of GSK4112 increased these calcium signals (Fig. 4B) that were statistically significant increased in the area under the curve (Fig. 4C). In some alphaTC1-9 cells that were silent at low glucose concentrations the addition of GSK4112 was able to trigger intracellular calcium oscillations (Fig. 4D). Similar results were achieved in mouse primary alpha-cells. Primary mouse alpha-cells exhibited the characteristic oscillatory calcium pattern at 0.5 mM glucose and addition of GSK4112 increased the signal frequency or transformed calcium oscillations into a sustained pattern (Fig. 4E–F). The synthetic Rev-erb alpha antagonist SR8278 had the opposite effect on glucagon secretion and intracellular calcium levels in alphaTC1-9 cells. At 0.5 mM glucose, SR8278 inhibited glucagon secretion and had no additional effect at 11 mM glucose (Fig. 5A). The effect of SR8278 on intracellular calcium was evident by its inhibitory action at 0.5 mM glucose (Fig. 5B and 5C). These results demonstrated that modulation or inhibition of Rev-erb alpha regulates glucagon secretion and alpha-cell calcium signalling.

Since AMPK has been shown to be strongly regulated by glucose in pancreatic alpha-cells, participates in glucagon release [16], and modulates the expression of different clock genes in skeletal muscle, we hypothesized that AMPK might be involved in the mechanism by which glucose regulates Rev-erb alpha gene expression. We treated alphaTC1-9 cells for 6 and 24 hours with 500 µM of the AMPK activator Metformin since the AMPK activator AICAR had no effect on alphaTC1-9 cells, as previously shown [16]. As expected, glucose down-regulated Rev-erb alpha gene expression at both 6 and 24 hours of treatment (Fig. 6A and B). Remarkably, while AMPK activation by metformin had no effect at 0.5 mM glucose, it prevented the inhibitory effect of 11 mM glucose on Rev-erb alpha gene expression (Fig. 6A and B) indicating that AMPK is involved in this process. Similar results were obtained with others clock genes such as Clock (Fig. 6C and D), Bmal1 (Fig. 6E and F) and Per2 (Fig. 6G and H).

It has been demonstrated in white adipose tissue that metformin can control clock genes expression through AMPK-Nampt-Sirt1 pathway [17]. To investigate whether a similar mechanism takes place in pancreatic alphaTC1-9 cells, we next studied this pathway by treating alphaTC1-9 cells for 6 and 24 hours with 500 µM Metformin. Figure 6 shows that 11 mM glucose inhibited Nampt and Sirt1 mRNA levels at 6 hours (Fig. 7A and C) and at 24 hours (Fig. 7B and D). Strikingly, metformin treatment reversed the inhibitory effect of glucose on Nampt (Fig. 6A and B) and Sirt1 (Fig. 7C and D). Consistent with the idea of a glucose-activated AMPK-Nampt-Sirt1 pathway, PGC-1 alpha (peroxisome proliferator-activated receptor gamma coactivator 1-alpha) gene expression (a target of AMPK and Sirt1) was also decreased at high glucose concentrations (Fig. 7E) and AMPK activation by metformin partially prevented this effect (Fig. 7E). The same effect on pAMPK in the Thr¹⁷² and Nampt protein levels were achieved in alpha alphaTC1-9 cells treated for 6 h (Fig. 7G and 7H) respectively.

To further examine the role of Nampt-Sirt1 signalling on Rev-erb alpha gene expression and glucagon secretion, we used FK866, a highly specific Nampt inhibitor. We treated alphaTC1-9 cells with 500 nM FK866 for 6 hours at low and high glucose concentrations and checked the mRNA expression of Sirt1 and PGC-1 alpha as downstream targets of Nampt. Figure 8A shows that Nampt inhibition by FK866 decreases Sirt1 expression at 0.5 mM glucose while no additional effect was seen at 11 mM glucose. PGC-1 alpha, the downstream target of Sirt1 was also decreased in the presence of FK866 at low glucose concentrations (Fig. 8B). Importantly, Nampt inhibition decreased Rev-erb alpha mRNA levels at 0.5 mM glucose whereas at 11 mM glucose when Nampt is already inhibited by glucose, (Fig. 7A and B) this effect was the same as glucose alone (Fig. 8C). Similar results were seen in Per2 mRNA levels in the presence of FK866 (Fig. 8D). Finally, we checked whether inhibition of Nampt affects glucagon secretion in isolated mouse pancreatic islets. As expected, glucose inhibited glucagon secretion at 11 mM glucose (Fig. 8E). Similarly, FK866 inhibitied glucagon secretion at 0.5 mM glucose with no additional effect at 11 mM glucose. These results demonstrate that glucose via AMPK-Nampt-Sirt1 pathway can control Rev-erb alpha gene expression and glucagon secretion in pancreatic alpha-cells. We therefore propose that at low glucose concentrations AMPK is activated which will trigger the Nampt-Sirt pathway increasing Rev-erb alpha expression levels and glucagon secretion. Conversely, high glucose will inhibit AMPK-Nampt-Sirt pathway and consequently Rev-erb alpha expression levels leading to a decrease in glucagon release (Fig. 8F).