Functional insulin receptor and its downstream effector PI3K (phosphoinositide 3-kinase) have been identified in pancreatic β-cells, but their involvement in the regulation of insulin secretion from β-cells remains unclear. In the present study, we investigated the physiological role of insulin and PI3K in glucose-induced biphasic insulin exocytosis in primary cultured β-cells and insulinoma Min6 cells using total internal reflection fluorescent microscopy. The pretreatment of β-cells with insulin induced the rapid increase in intracellular Ca2+ levels and accelerated the exocytotic response without affecting the second-phase insulin secretion. The inhibition of PI3K not only abolished the insulin-induced rapid development of the exocytotic response, but also potentiated the second-phase insulin secretion. The rapid development of Ca2+ and accelerated exocytotic response induced by insulin were accompanied by the translocation of the Ca2+-permeable channel TrpV2 (transient receptor potential V2) in a PI3K-dependent manner. Inhibition of TrpV2 by the selective blocker tranilast, or the expression of shRNA (short-hairpin RNA) against TrpV2 suppressed the effect of insulin in the first phase, but the second phase was not affected. Thus our results demonstrate that insulin treatment induced the acceleration of the exocytotic response during the glucose-induced first-phase response by the insertion of TrpV2 into the plasma membrane in a PI3K-dependent manner.
Insulin is stored in large dense-core granules in pancreatic β-cells and is released by Ca2+-dependent exocytosis in response to an elevation of the blood glucose level. In pancreatic β-cells, glucose stimulation evokes an increase in the ATP/ADP ratio, which triggers the closure of ATP-sensitive K+ channels (KATP channels) and cell depolarization. Activation of voltage-gated Ca2+ channels by membrane depolarization increases Ca2+ influx and stimulates insulin secretion [1,2]. The insulin secretion from pancreatic β-cells shows a characteristic biphasic pattern consisting of a rapidly developing and transient first phase followed by a sustained second phase [3,4], and we and other groups showed that the first and second phases of insulin secretion were controlled by distinct molecular mechanisms [5,6].
Secreted insulin functions through the IR (insulin receptor) and activates downstream targets. PI3K (phosphoinositide 3-kinase) is a key component that transmits the insulin signal to downstream effectors, and has also been reported to regulate the trafficking of intracellular vesicles and exocytosis in many cell types [7,8]. Because functional IR and PI3K are expressed in pancreatic β-cells [9–11], an autocrine effect of insulin on its own secretion has been postulated. Several studies have demonstrated positive feedback regulation of insulin in the secretory function of β-cells, e.g. the first-phase insulin secretion was suppressed in β-cells lacking IR  and the pretreatment of β-cells with anti-insulin antibody inhibited glucose-induced insulin secretion . On the other hand, other studies have shown that insulin negatively regulates insulin secretion from pancreatic β-cells. For example, the genetic ablation of p85α, a major regulatory subunit of PI3K in β-cells, resulted in the enhancement of insulin secretion preferentially during the second phase . Furthermore, a highly selective inhibitor of PI3K, wortmannin, was also reported to enhance insulin secretion [15–17], thus indicating that the insulin/PI3K pathway functions as a negative regulator of insulin secretion. In addition, studies involving perfusion of the pancreas or perifusion of isolated islets have demonstrated that exogenous insulin inhibited C-peptide or endogenous insulin secretion [18–24], although other groups have disputed these results [25–27]. Taken together, the results suggest that insulin could control multiple downstream effectors, each of which could positively or negatively regulate insulin secretion. Therefore, in order to understand the autocrine function of insulin, it is necessary to evaluate the role of each downstream effector activated by the insulin/PI3K pathway in biphasic insulin secretion.
Recently, the Ca2+-permeable cation channel TrpV2 (transient receptor potential V2) was reported to be inserted into the plasma membrane in response to PI3K activation . TrpV2 is a member of the transient receptor potential channel family that is well-recognized as unique cellular sensors for mechano- and thermo-stimulations [28–30]. Noxious heat, osmotic and mechanical stimulations, and some molecules have been reported to activate TrpV2 [30–34], but the endogenous ligand(s) is unknown. In addition, the physiological role of TrpV2 in glucose-induced biphasic insulin secretion remains to be elucidated, although TrpV2 has been shown to be involved in the regulation of insulin secretion .
In the present study, we investigated the role of the insulin/PI3K pathway in insulin secretion, using primary cultured β-cells and the insulinoma cell line Min6, by TIRF (total internal reflection fluorescent) microscopy-based imaging analysis. TIRF microscopy together with the expression of GFP (green fluorescent protein)-tagged insulin enabled us to detect the exocytosis of individual insulin granules with high time-resolution regardless of the presence of endogenous insulin. Our results demonstrated that insulin accelerated the onset of the exocytotic response to the glucose stimulation via the insertion of TrpV2 into the plasma membrane. On the other hand, insulin negatively regulated the fusion of insulin granules during the second phase.
Plasmids and adenoviruses
cDNA encoding a full open reading frame of the mouse TrpV2 was amplified from total RNA isolated from Min6 cells by RT (reverse transcriptase)–PCR using specific primers (5′-CCACAGAAGTTTCAGCGATAAGGAGCACCCTC-3′ and 5′-GCAAAATTCCCTACTCTACCCTGCCAGCCTG-3′). The purified PCR product was directly inserted into the pGEM-T easy vector (Promega). The TrpV2 sequence, without the stop codon, was amplified by PCR and subcloned into the pENTR1A vector (Invitrogen). A DNA fragment corresponding to the red florescent protein mCherry, a gift from Dr Haruhiko Bito (University of Tokyo, Tokyo, Japan), was also subcloned into pENTR1A. The DN (dominant-negative) form of the human p85α [p85α(DN)] was amplified by PCR from the recombinant adenovirus  and subcloned into pENTR1A with mCherry. To generate the mammalian expression vectors for V5-tagged TrpV2 and mCherry, these sequences in pENTR1A were introduced into the pcDNA3.2/V5-DEST vector (Invitrogen) by site-specific recombination according to the manufacturer's instructions. To generate the adenovirus vector, p85α(DN) tagged with mCherry and mCherry alone in pENTR1A were introduced into the pAd/CMV/V5-DEST vector (Invitrogen) by site-specific recombination. The recombinant adenoviruses that express p85α(DN)–mCherry or mCherry alone were generated according to the manufacturer's instructions. An expression vector for the shRNA (short-hairpin RNA) against the mouse TrpV2 was generated by the ligation of annealed DNA fragments (5′-GATCCGTGGCTGAACCTGCTTTATTCTCAAGAGAAATAAAGCAGGTTCAGCCAGCTTTTTTGGAAA-3′ and 5′-AGCTTTTCCAAAAAAGCTGGCTGAACCTGCTTTATTTCTCTTGAGAATAAAGCAGGTTCAGCCACG-3′) into the pSilencer vector (Ambion). A vector with the scrambled shRNA sequence (Ambion) was used as a negative control. All constructs were fully sequenced to confirm their identity.
Cell culture, infection and transfection
Pancreatic islets of Langerhans were isolated from male C57BL/6 mice by collagenase digestion as described previously . Briefly, isolated islets were dispersed in calcium-free KRB (Krebs–Ringer buffer) containing 1 mM EGTA and cultured on fibronectin-coated high-refractive-index coverslips (Olympus) in RPMI 1640 medium (Invitrogen) supplemented with 10% FBS (fetal bovine serum; Invitrogen), 200 units/ml penicillin and 200 μg/ml streptomycin at 37 °C under a 5% CO2 atmosphere. To label the insulin secretory granules and express p85α(DN)–mCherry or mCherry alone, cultured β-cells were infected with recombinant adenoviruses Adex1CA insulin–GFP and pAd/CMV/p85α(DN)-mCherry or pAd/CMV/mCherry simultaneously. After 18–24 h, the cells were used for the experiments.
Min6 cells, a gift from Dr J. Miyazaki (Osaka University, Kanan, Japan), at passage 20–30 were cultured as described previously . Min6 cells were plated on to fibronectin-coated high-refractive-index glass for imaging by TIRF microscopy, on to fibronectin-coated normal coverslips for immunostaining and Ca2+ imaging or on to fibronectin-coated 60-mm-diameter dishes dishes for biochemical analysis. After 18–24 h, cells were transfected with various vectors using Lipofectamine™ 2000 (Invitrogen) in the presence of serum according to the manufacturer's instructions. Culture medium was changed from DMEM (Dulbecco's modified Eagle's medium) containing 11 mM glucose to DMEM with 5 mM glucose at 18–24 h before the experiments. Experiments were performed within 3 days after plating.
The Olympus total internal reflection system was used with a high-aperture objective lens (Apo 100× OHR, numerical aperture 1.65; Olympus) essentially as described previously [5,37]. To observe the fluorescence of GFP, we used a 488-nm laser line for excitation and a 515-nm long-pass filter for the barrier. Images were projected on to a cooled CCD (charge-coupled device) camera (DU-897E; Andor Technology) operated with Metamorph version 7.5 (Universal Imaging). Images were acquired at 300-ms intervals. Diiodomethane sulfur immersion oil (n=1.81; Cargille Laboratories) was used to make contact between the objective lens and the high-refractive-index coverslip. Light propagates through the coverslip at an angle of 65 ° at which the predicted penetration depth should be approx. 44 nm . For real-time imaging of the insulin granules labelled by insulin–GFP under TIRF microscopy, β-cells were incubated for 20 min at 37 °C in KRB containing 10 mM Hepes, pH 7.4, 110 mM NaCl, 4.4 mM KCl, 1.45 mM KH2PO4, 1.2 mM MgSO4, 2.3 mM calcium gluconate, 4.8 mM NaHCO3, 4 mM glucose and 0.3% BSA. The cells were then transferred on to the stage of TIRF microscope, and stimulation with glucose was achieved by the addition of 40 mM glucose in KRB into the chamber (to give a final concentration of 22 mM glucose). For the imaging of GFP in cells expressing mCherry, we used light from a Xenon lamp passed through a 535±15-nm band-pass filter for excitation and a 580-nm long-pass filter for the barrier. Cells with the mCherry signal were selected, then GFP signals were monitored as described above. Data analysis was performed using Metamorph software. Fusion events were manually counted while looping at a 3500 frame time-lapse. Results are presented as means±S.E.M.
Fura 2 fluorimetry
Min6 cells were labelled with 10 μM fura 2/AM (fura 2 acetoxymethyl ester; Invitrogen) for 20 min at 37 °C in KRB with 4 mM glucose, washed and incubated for an additional 20 min with KRB. Coverslips were mounted on to an ARGUS/HiSCA system (Hamamatsu Photonics). Fura 2 fluorescence was detected by the cooled CCD camera after excitation at 340 nm (F340) and 380 nm (F380), and the ratio of the image (F340/F380) was calculated with the ARGUS/HiSCA system. For Ca2+ imaging in cells expressing shRNA and mCherry, we used light from a Xenon lamp passed through a 535±15-nm band-pass filter for excitation and a 580-nm long-pass filter for the barrier. Cells with the mCherry signal were selected, then fura 2 signals were monitored as described above.
hGH (human growth hormone) release assay
hGH release was determined essentially as described previously . Briefly, Min6 cells expressing hGH were washed three times with KRB containing 4 mM glucose. After washing, cells were incubated with KRB containing 4 mM glucose for 15 min, and then stimulated with KRB containing 4 or 22 mM glucose for 3 or 15 min. After the end of the experiments, 1 ml of chilled 1% (v/v) Nonidet P-40 was added and then the samples were sonicated (for 30 s; UR-20P, Tomy Seiko) on ice. Secreted hGH and the total cellular content of hGH were measured with a hGH ELISA kit (Roche).
Min6 cells cultured on coverslips were fixed and permeabilized with 4% (w/v) paraformaldehyde/0.1% Triton X-100 and were processed for immunocytochemistry as described previously . Cells were labelled with rabbit anti-TrpV2 polyclonal antibody (at 1:200; Biomol) and processed with goat anti-(rabbit IgG) conjugated with Alexa Fluor® 488 (Invitrogen). Immunofluorescence was detected by a laser-scanning confocal system (FV1000; Olympus), and signal intensity per μm2 was determined using Metamorph software.
Purification of cell-surface proteins by biotinylation and immunobloting
Min6 cells grown on 60-mm-diameter dishes were washed twice with KRB containing 4 mM glucose, and incubated in KRB with or without 100 ng/ml insulin and 100 nM wortmannin for 20 min at 37 °C. After the incubation, proteins expressed in the plasma membrane were labelled by sulfo-NHS (N-hydroxysuccinimido)-biotin (Pierce) and isolated by NeutrAvidin beads according to the instructions provided.
SDS/PAGE was performed on 8% polyacrylamide gels equipped with a 5% stacking gel. After separation, the proteins were transferred on to PVDF membranes following standard procedures with a semi-dry transblotting apparatus. The membranes were blocked in 5% (w/v) non-fat milk powder in TBST (Tris-buffered saline with Tween 20; 25 mM Tris/HCl, pH 7.5, 150 mM NaCl and 0.05% Tween 20) for 30 min at room temperature (22 °C) followed by an overnight incubation with anti-TrpV2 (1:200), anti-α1C (1:200; Alomone Labs), anti-V5 (1:5000; Invitrogen) and anti-α-tubulin (1:5000; Sigma–Aldrich) antibodies diluted in 5% (w/v) non-fat milk powder in TBST at 4 °C. After washing with TBST, membranes were incubated for 1 h at room temperature with horseradish peroxidase-labelled anti-(rabbit IgG) or anti-(mouse IgG) antibody (1:1000; Dako) in TBST containing 5% (w/v) non-fat milk powder. After washing, the immunoreactive bands were visualized by SuperSignal (Pierce) and a luminescence image analyser with an electronically cooled CCD camera system (LAS-3000; Fuji Photo Film).
Insulin bi-directionally regulates first- and second-phase insulin secretion via PI3K
We examined whether insulin could activate the insulin signalling pathway in primary cultured pancreatic β-cells. Cells were incubated in KRB with or without 100 ng/ml insulin (~17 nM) for 20 min. The phosphorylation of Akt, a downstream target of PI3K , was markedly enhanced by insulin treatment (results not shown), indicating that the insulin/PI3K pathway was not saturated in our culture condition. In order to examine the effect of insulin on glucose-induced insulin secretion, we first utilized the β-cell-derived clonal cell line Min6, transfected with hGH, which serves as a marker for exocytosis  and enables us to distinguish the exocytotic activity from exogenously added insulin. Min6 cells transfected with hGH were stimulated with 4 or 22 mM glucose, and secreted hGH was measured by ELISA. As shown in Figure 1, the hGH secretion evoked by a 3-min exposure to 22 mM glucose was significantly enhanced by insulin treatment (0.27±0.02% and 0.47±0.06% of total cellular content in control and insulin-treated Min6 cells respectively; Figure 1A). However, when cells were stimulated with 22 mM glucose for 15 min, the pretreatment with insulin did not potentiate the hGH secretion (1.46±0.11% and 1.52±0.15% of total cellular content in control and insulin-treated Min6 cells respectively; Figure 1B). In addition, the effect of insulin treatment on a 3-min stimulation was blocked by the highly selective PI3K inhibitor wortmannin (0.12% ±0.03, 0.39% ±0.06 and 0.18±0.05% of total cellular content in control, insulin-, and insulin/wortmannin-treated Min6 cells respectively, Figure 1C). These results suggested that insulin might have an autocrine effect on the beginning of the first-phase, but not the second-phase, of insulin secretion in a PI3K-dependent manner.
Insulin treatment enhances the glucose-stimulated first-phase, but not the second-phase, secretion in Min6 cells
To investigate further the autocrine effect of insulin on glucose-induced exocytosis, we took advantage of TIRF microscopy that illuminates a very thin layer immediately adjacent to the coverslips [41,42]. This allows the real-time imaging of glucose-induced exocytosis of insulin granules in the presence of exogenously added insulin in β-cells expressing insulin–GFP. Primary cultured β-cells expressing insulin–GFP were incubated with or without 100 ng/ml insulin for 15 min, then the motion of insulin granules was observed under TIRF microscopy. We found that acute insulin treatment did not affect the number of docked granules on the plasma membrane as detected by TIRF microscopy (99.8±11.8 and 93.9±7.8 events per 200 μm2 for control and insulin-treated β-cells respectively). When β-cells expressing insulin–GFP were stimulated with 22 mM glucose, we detected exocytotic responses that originated from two distinct types of insulin granules with different behaviours prior to fusion. As we reported previously [5,36,37], we defined fusions originating from insulin granules that are visible before the onset of stimulation under TIRF microscopy as ‘previously docked granules’. On the other hand, fusions also arose from insulin granules that could not be detected by TIRF microscopy, or were dimly visible before stimulation, and that fuse immediately after they reach the plasma membrane; these are defined as ‘newcomer granules’. As shown in Figure 2(A), the number of exocytotic responses per minute during the first-phase glucose secretion (<7 min) gradually increased and peaked at 5.38±0.60 min after the onset of 22 mM glucose stimulation. On the other hand, in β-cells pretreated with insulin for 15 min, the number of exocytotic events steeply increased after ~2 min latency from the onset of the stimulation, and the highest peak was detected at 3.80±0.32 min (Figure 2B). This effect of insulin was highly likely to depend on the PI3K activity because two PI3K inhibitors with distinct molecular structures, 100 nM wortmannin and 1 μM PIK-75 , completely blocked the insulin-induced rapid development of the exocytotic response (the peaks were at 5.30±0.62 and 6.00±0.52 min for insulin and wortmannin- or PIK-75-treated β-cells respectively; Figures 2C–2E). These results were compatible with those obtained using standard biochemical techniques in Min6 cells (Figures 1A and 1C). Total fusion events detected during the first phase (<7 min) were considerably increased by insulin treatment (43.0±4.9 and 53.0±6.0 events per 200 μm2 for control and insulin-treated cells respectively), and PI3K inhibitors suppressed the insulin-induced increase in exocytotic events (39.8±4.9 and 45.5±3.5 events per 200 μm2 for insulin/wortmannin- or insulin/PIK-75-treated β-cells respectively; Figure 2F). On the other hand, insulin treatment did not affect the number of exocytotic events detected during the second-phase glucose secretion (>7 min) (36.0±6.4 and 42.3±6.7 events per 200 μm2 for control and insulin-treated cells respectively). However, PI3K inhibitors markedly enhanced the exocytotic responses (61.6±5.5 and 81.7±12.0 events per 200 μm2 for insulin/wortmannin-treated or insulin/PIK-75-treated β-cells respectively, Figure 2G).
Insulin treatment induces the rapid development of the exocytotic response via PI3K in primary cultured pancreatic β-cells
To examine further the participation of PI3K in the effects of insulin on the glucose-induced exocytotic response, we next studied the effects of inhibition of PI3K activity using a DN form of p85α [p85α(DN)]. As shown in Figure 3, the expression of p85α(DN)–mCherry suppressed the effect of insulin on the rapid development of the exocytotic response, whereas the expression of mCherry alone did not alter the effect of insulin (the peaks were at 4.86±0.34, 3.83±0.31, 4.64±0.34 min for control and insulin-treated β-cells expressing mCherry, and insulin-treated β-cells expressing p85α(DN)–mCherry respectively). In addition, the number of total fusion events observed during the second phase of insulin secretion was markedly enhanced in β-cells expressing p85α(DN) (Figure 3F), as observed in PI3K-inhibitor-treated cells. These results indicate that insulin induces two distinct effects on the biphasic exocytotic response in pancreatic β-cells: (i) the rapid development of the first-phase exocytotic response; and (ii) the negative regulation of the second-phase exocytosis.
Expression of p85α(DN) blocked the rapid development of the exocytotic response induced by insulin
Insulin modulates the first-phase [Ca2+]i influx but not second-phase [Ca2+]i dynamics
To examine whether the effects of insulin on the first- and second-phase exocytosis were related to a change in Ca2+ dynamics induced by glucose stimulation, we next conducted [Ca2+]i measurement using fura 2 microfluorimetry in primary cultured β-cells. As shown in Figure 4(A), insulin induced a rapid increase in the [Ca2+]i in response to 22 mM glucose. The peak [Ca2+]i was detected at 2.85±0.13 and 2.07±0.09 min after the onset of stimulation in control and insulin-treated β-cells respectively (Figure 4B). This effect of insulin on the rapid development of the [Ca2+]i response was inhibited by PI3K inhibitors ([Ca2+]i peak at 2.72±0.09 and 3.16±0.12 min after the onset of stimulation in insulin/wortmannin- or insulin/PIK-75-treated β-cells respectively). On the other hand, insulin did not induce any significant change in [Ca2+]i at >7 min after the onset of the stimulation. These results suggest that the insulin-induced rapid development of the Ca2+ response caused the rapid exocytotic response, whereas insulin negatively regulated the second-phase insulin secretion without affecting intracellular Ca2+ mobilization.
Effect of insulin on the glucose-induced [Ca2+]i response
Insulin recruits TrpV2 to the plasma membrane via PI3K activation
From the results described above, it was conceivable that the insulin-induced rapid development of [Ca2+]i could lead to accelerated exocytotic responses to glucose stimulation. It has been reported that the translocation of TrpV2, a Ca2+-permeable channel, from the intracellular pool to the plasma membrane is regulated by PI3K [28,29]. Indeed, we detected TrpV2 immunoreactivity in the plasma membrane that was enhanced by a 15 min treatment with 100 ng/ml insulin (Figures 5A and 5B). To further evaluate the translocation of TrpV2 from the intracellular pool to the plasma membrane by insulin treatment, we biochemically examined the amount of TrpV2 on the plasma membrane. After treatment of Min6 cells with insulin, membrane proteins exposed to the extracellular surface were labelled using a membrane-impermeant biotinylation reagent, and biotinylated proteins were recovered by NeutrAvidin beads. As shown in Figure 5(C), insulin increased the amount of biotinylated TrpV2 by up to 70% compared with the control condition. Wortmannin completely inhibited the increase in biotinylated TrpV2 induced by insulin, indicating that insulin induced the insertion of TrpV2 into the plasma membrane in a PI3K-dependent manner. It was of particular note that insulin did not affect the amount of α1C, a pore-forming subunit of the L-type Ca2+ channel, on the plasma membrane (Figure 5D). These results indicate that insulin recruited TrpV2, but not L-type Ca2+ channels, to the plasma membrane via PI3K activation.
Insulin induces the translocation of TrpV2 to the plasma membrane
Effects of TrpV2 blockage on insulin secretion
Since insulin induced the rapid development of the Ca2+ response and the recruitment of TrpV2 to the plasma membrane, we expected that the insulin-induced recruitment of TrpV2 would result in additional Ca2+ influx through TrpV2, which could cause the rapid development of Ca2+ and subsequently accelerate the exocytotic response to glucose stimulation. To assess this possibility, we next studied the effect of the selective TrpV2 blocker tranilast [29,44,45] on the exocytotic response in insulin-treated β-cells. As shown in Figure 6(A), tranilast inhibited the insulin-induced rapid development of the [Ca2+]i response during the first-phase response. The peaks of the [Ca2+]i elevation were observed at 3.39±0.14, 2.19±0.11 and 4.06±0.14 min after the onset of glucose stimulation in control, insulin- and insulin/tranilast-treated cells respectively (Figure 6B). On the other hand, during the second-phase response (>7 min), tranilast only slightly reduced [Ca2+]i, and the difference was not statistically significant.
Effect of tranilast on the insulin-induced rapid development of [Ca2+]i response
In agreement with these [Ca2+]i fluorimetric results, tranilast also abolished the insulin-induced rapid development of the exocytotic response (Figure 7). The peaks of fusion events were observed at 5.77±0.32, 4.00±0.26 and 6.00±0.27 min after the onset of glucose stimulation in control, insulin- and insulin/tranilast-treated β-cells respectively (Figure 7D). The insulin-induced increase in the total number of fusion events during the first-phase response was also blocked by tranilast treatment (Figure 7E). Interestingly, unlike with PI3K inhibitors, tranilast did not enhance the exocytotic response during the second-phase response (Figure 7F). Thus the negative regulation of the second-phase insulin secretion by insulin was independent of TrpV2.
Effect of tranilast on the insulin-induced rapid development of exocytotic response
To rule out the possibility of a non-specific effect of tranilast, we attempted to reduce the expression level of TrpV2 using shRNA. Because of the very poor transfection efficacy in primary cultured β-cells, we used Min6 cells in these experiments. The shRNA against TrpV2 almost completely blocked the exogenously introduced TrpV2 expression (Figure 8A) and effectively reduced the amount of endogenous TrpV2 (Figures 8B and 8C). To examine the effect of TrpV2 shRNA on the glucose-induced [Ca2+]i response, Min6 cells were transfected with TrpV2 shRNA and mCherry and the Ca2+ response was assessed by fura 2 fluorimetry in cells expressing mCherry. As shown in Figure 9(A), the expression of TrpV2 shRNA impaired the insulin-induced rapid development of the Ca2+ response ([Ca2+]i peak at 4.36±0.33, 2.10±0.33 and 3.47±0.22 min after the onset of glucose stimulation in control, insulin- and insulin/TrpV2 shRNA-treated Min6 cells respectively). The expression of shRNA against TrpV2 also abolished the insulin-induced rapid development of the exocytotic response (Figure 10). The peaks of the exocytotic responses were detected at 4.83±0.48, 2.44±0.24 and 6.00±0.72 min after the onset of stimulation in control, insulin- and insulin/TrpV2 shRNA-treated cells respectively (Figure 10D). Therefore the reduction in TrpV2 inhibited the effect of insulin on the rapid elevation of [Ca2+]i and accelerated the exocytotic response during the first-phase response.
shRNA against TrpV2 suppressed the expression of TrpV2 in Min6 cells
Taken together, our results indicate that the insulin-induced recruitment of TrpV2 to the plasma membrane is involved in the rapid development of the Ca2+ and exocytotic response during the first-phase, but not the second-phase, exocytotic response.
In the present study, we found that insulin induced the rapid development of the exocytotic response to glucose stimulation in a PI3K-dependent manner, and that this was partially achieved by the PI3K-dependent translocation of TrpV2 from the intracellular pool to the plasma membrane.
Although a large number of studies have been published about the autocrine effect of insulin on its own secretion [18–27], most of them were based on the biochemical detection of secreted insulin or C-peptide from pancreatic β-cells, which shows poor temporal resolution. Amperometric recording in 5-HT (5-hydroxytryptamine, also known as serotonin)-loaded cells enabled us to detect the exocytotic response in the presence of exogenous insulin , but the technique cannot distinguish the exocytosis of insulin granules from that of other secretory granules. In addition, a report showing that 5-HT affected insulin secretion in pancreatic β-cells raised questions about the results from 5-HT-loaded cells . On the other hand, our imaging analysis based on TIRF microscopy enabled us to detect individual exocytotic responses from insulin granules in the presence of exogenous insulin with high time-resolution [5,36,37], which allowed us to address the issue of the autocrine effect of insulin on its own secretion.
The insulin-induced rapid development of the Ca2+ response observed in the present study was in good agreement with a report showing that the onset of the glucose-induced Ca2+ response was delayed in β-cells lacking IRS (insulin receptor substrate)-1 . Since it was reported that acute inhibition of PI3K does not affect glucose metabolism in pancreatic β-cells , insulin signalling through PI3K activation should modulate the step(s) subsequent to the depolarization evoked by the closure of the KATP channel. Since PI3K has been reported to regulate the number of voltage-gated Ca2+ channels in the plasma membrane in neurons and cardiac myocytes [48,49] and that of TrpV2 in pancreatic β-cells , we examined the PI3K-dependent insertion of these two Ca2+-permeable channels. We found that insulin induced the recruitment of TrpV2, but not voltage-gated Ca2+ channels, from the intracellular pool to the plasma membrane in Min6 cells (Figure 5). Consistent with this biochemical result, both the inhibition of TrpV2 by tranilast and the reduction of TrpV2 by shRNA against TrpV2 blocked the insulin-induced rapid development of the [Ca2+]i response (Figures 6 and 9), indicating that the PI3K-dependent insertion of TrpV2 into the plasma membrane is responsible for the effect of insulin on glucose-induced [Ca2+]i mobilization. Although Hisanaga et al.  showed that the glucose-evoked [Ca2+]i during the second-phase, but not the first-phase, response was affected in Min6 cells expressing shRNA against TrpV2, they did not confirm whether the cells they observed had been successfully transfected. Therefore it might be possible that their result could include data from untransfected cells. On the other hand, in the present study, since shRNA against TrpV2 and mCherry were co-transfected, we could analyse the [Ca2+]i response in cells with the mCherry signal without contamination with the results from untransfected cells. From these experiments, we concluded that the insulin-induced TrpV2 recruitment to the plasma membrane accelerated the development of the [Ca2+]i observed during the first-phase response.
Hisanaga et al.  used a standard biochemical approach to show that TrpV2 was involved in the Ca2+-dependent insulin secretion. They assumed that TrpV2 may affect the second-phase insulin secretion because their Ca2+ imaging in Min6 cells showed that TrpV2 slightly affected the Ca2+ influx during the second-phase response. However, their study did not clarify which phase of the glucose-induced biphasic insulin secretion was influenced by TrpV2 in β-cells. In the present study, in order to address the physiological role of TrpV2 on biphasic insulin secretion, we examined the effect of blockage of TrpV2 function on glucose-induced biphasic secretion by real-time imaging using TIRF microscopy. We have demonstrated that acute treatment with tranilast and the expression of shRNA against TrpV2 caused the suppression of the first-phase exocytotic responses in Min6 and primary cultured β-cells. In contrast, we showed that TrpV2 was not likely to be involved in the second-phase insulin secretion (Figures 7 and 10). These results were in good agreement with those from our Ca2+ imaging (Figures 6 and 9). Thus our present study clearly demonstrates that TrpV2 regulated the glucose-induced first-phase, rather than second-phase, insulin secretion.
Effect of TrpV2 shRNA on the insulin-induced rapid development of the [Ca2+]i response in Min6 cells
Effect of TrpV2 shRNA on the insulin-induced rapid development of the exocytotic response in Min6 cells
The endoplasmic reticulum is another important source of Ca2+ that triggers the exocytosis of insulin . Aspinwall et al. [51,52] reported that a high concentration of insulin (100 nM) induced Ca2+ efflux from the intracellular Ca2+ pool and, consequently, insulin secretion. However, the insulin-induced rapid development of the [Ca2+]i and exocytotic responses observed in the present study would not have been caused by the insulin-induced Ca2+ efflux from the intracellular Ca2+ pool, because the concentration of insulin used (100 ng/ml, ~17 nM) was not high enough to evoke that response [51,52]. On the other hand, 10 nM insulin was sufficient to recruit TrpV2 to the plasma membrane . All of these results indicate that insulin treatment induces the translocation of TrpV2 to the plasma membrane, resulting in the rapid development of the [Ca2+]i and exocytotic responses.
The results from the present study have shown that TrpV2 is involved in the development of the [Ca2+]i response upon glucose stimulation in β-cells, but its molecular mechanism is still unclear. To date, noxious heat, mechanical stimulation and some molecules have been reported to activate TrpV2 [30–34], and PI3K and Ca2+/calmodulin-dependent kinase have been demonstrated to regulate the Ca2+ current through TrpV2 [28,53,54]. The biochemical analysis in the present study demonstrated that the amount of TrpV2 in the plasma membrane was increased up to 70% by insulin treatment, but the basal [Ca2+]i was not affected in insulin-treated Min6 cells (Figure 5). Therefore it is reasonable that TrpV2 inserted into the plasma membrane was not active in the basal condition, but opened in response to glucose stimulation. The change in the local membrane tension by the fusion of secretory granules may trigger the TrpV2 activation. Alternatively, Ca2+ influx through voltage-gated Ca2+ channels may activate TrpV2 (given a Ca2+ channel blocker has been reported to inhibit Ca2+ influx through TrpV2 ).
In the present study, we showed the positive and negative effects of insulin for the first- and second-phase insulin secretion respectively (Figures 2 and 3). Because insulin and PI3K inhibitors did not affect [Ca2+]i during the second-phase response, PI3K should negatively regulate the exocytotic process subsequent to Ca2+ influx during the second phase. This is in good agreement with a previous report showing that genetic ablation of p85α resulted in the potentiation of insulin secretion preferentially during the second phase . Interestingly, the inhibition of PI3K selectively enhanced exocytotic responses that originated from newcomer rather than previously docked granules (Figures 2 and 3). Because the exocytosis of newcomer granules is syntaxin 1A-independent, it is of note that the molecular mechanism underlying the exocytosis of newcomer granules should be different from that of previously docked granules . Thus further study to clarify the molecular mechanism of the PI3K-dependent negative regulation of the second-phase insulin release would also lead to the elucidation of the molecular mechanism involved in the fusion of newcomer granules.
Although we have presented evidence that insulin affected the responsiveness of β-cells to glucose stimulation in a PI3K-dependent manner, it is uncertain at present whether this autocrine effect of insulin actually controls translocation of TrpV2 in vivo. It is possible that basal insulin secretion would be sufficient to recruit the maximum number of TrpV2 channels into the plasma membrane. On the other hand, it might be also possible that the amount of TrpV2 in the plasma membrane is reduced in β-cells at an initial phase of diabetes, resulting in a poor first-phase exocytotic response to glucose stimulation. In fact, such a delayed Ca2+ response to glucose stimulation has been reported in β-cells prepared from diabetic GK (Goto–Kakizaki) rats .
In conclusion, insulin acts through bi-directional regulation in insulin exocytosis, namely acceleration of the onset of insulin secretion during the first-phase response and inhibition of insulin release during the second-phase response. The effect of insulin on the first-phase response was achieved by the recruitment of TrpV2 from the intracellular pool to the plasma membrane in response to the activation of PI3K. Further studies are necessary to assess the in vivo role of TrpV2 during the first phase and the molecular mechanism underlying the inhibitory effect of the insulin/PI3K pathway on insulin secretion during the second phase.
We greatly appreciate the gift of the recombinant adenovirus for p85α(DN) expression from Dr K. Ueki, mCherry vector from H. Bito and Min6 cells from Dr J. Miyazaki, and thank Mr Brent Bell (Department of Respiratory and Infectious Diseases, Tohoku University Graduate School of Medicine, Sendai, Japan) for critical reading of the manuscript prior to submission.
Dulbecco's modified Eagle's medium
green fluorescent protein
human growth hormone
Tris-buffered saline with Tween 20
total internal reflection fluorescent
transient receptor potential V2
Kyota Aoyagi, Chiyono Nishiwaki, and Yoko Nakamichi performed all of the experiments. Mica Ohara-imaizumi and Shinya Nagamatsu designed the experiments. Kyota Aoyagi, Mica Ohara-imaizumi and Shinya Nagamatsu co-wrote the manuscript.
This work was supported by KAKENHI [grant numbers C-20570189, 21113523, B-20390260], the Sumitomo Foundation; the Astellas Foundation for Research on Metabolic Disorders; the Research Foundation for Opto-Science and Technology; a Kyorin University School of Medicince Collaboration Project; and the Novartis Foundation (Japan) for the Promotion of Science.