SEARCH

SEARCH BY CITATION

Keywords:

  • glucosensing;
  • hypothalamus;
  • glucokinase;
  • connexons

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. REFERENCES
  8. Supporting Information

The ventromedial hypothalamus is involved in regulating feeding and satiety behavior, and its neurons interact with specialized ependymal-glial cells, termed tanycytes. The latter express glucose-sensing proteins, including glucose transporter 2, glucokinase, and ATP-sensitive K+ (KATP) channels, suggesting their involvement in hypothalamic glucosensing. Here, the transduction mechanism involved in the glucose-induced rise of intracellular free Ca2+ concentration ([Ca2+]i) in cultured β-tanycytes was examined. Fura-2AM time-lapse fluorescence images revealed that glucose increases the intracellular Ca2+ signal in a concentration-dependent manner. Glucose transportation, primarily via glucose transporters, and metabolism via anaerobic glycolysis increased connexin 43 (Cx43) hemichannel activity, evaluated by ethidium uptake and whole cell patch clamp recordings, through a KATP channel-dependent pathway. Consequently, ATP export to the extracellular milieu was enhanced, resulting in activation of purinergic P2Y1 receptors followed by inositol trisphosphate receptor activation and Ca2+ release from intracellular stores. The present study identifies the mechanism by which glucose increases [Ca2+]i in tanycytes. It also establishes that Cx43 hemichannels can be rapidly activated under physiological conditions by the sequential activation of glucosensing proteins in normal tanycytes. © 2011 Wiley Periodicals, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. REFERENCES
  8. Supporting Information

The ventromedial hypothalamus (VMH) is involved in regulating feeding and satiety behaviors through their capacity to detect changes in glucose concentrations (Levin et al.,2004). The arcuate nucleus (AN) and the ventromedial nucleus (VMN) form the VMH; their neurons are in close contact with highly elongated ependymal-glial cells known as tanycytes (Akmayev and Popov,1977; Chauvet et al.,1995; Flament-Durand and Brion,1985). Tanycytes are the main glial cells present in the basal hypothalamus (García et al.,2001, 2003; Millan et al.,2010) and are classified into four types, according to their localization in the III–V ventricle and biochemical and molecular properties: α1, α2, β1, and β2 (Akmayev and Fidelina,1974; Akmayev and Popov,1977). α1- and α2-tanycytes are localized beside the VMN, while β1-tanycytes are localized within the lower lateral wall of the third ventricle, contacting neurons through cell processes in the AN as well as capillaries in the hypothalamus (García et al.,2001). β2-tanycytes are found in the floor of third ventricle lining the median eminence (García et al.,2001).

Both α and β tanycytes express several glucose-sensing molecules, including glucose transporter 2 (GLUT2) and ATP-sensitive K+ (KATP) channels, suggesting their possible involvement in hypothalamus-mediated glucosensing (Alvarez et al.,1996; García et al.,2003; Millan et al.,2010; Navarro et al.,1996). Notably, unlike α-tanycytes, high expression of GLUT2 and glucokinase (GK) in the proximal pole has been observed in β1-tanycytes in vivo (García et al.,2003; Millan et al.,2010). The specific localization of β1 tanycytes in direct contact with cerebral spinal fluid and their prominent GLUT2/GK expression strongly support the idea that these cells have a high glucose uptake capacity. In fact, it has been proposed that β1-tanycytes could uptake and metabolize glucose to lactate through the glycolytic pathway, and subsequently export lactate to neurons of the AN through monocarboxylate transporters (MCTs) 1 and/or 4 (Cortés-Campos et al.,2011; García et al.,2003; Millan et al.,2010).

In support of the putative brain glucosensor role of tanycytes, it was recently demonstrated that extracellular glucose increases the intracellular free Ca2+ concentration ([Ca2+]i) in α1- and α2-tanycytes (Frayling et al.,2011). Although expression of glucosensing proteins by different tanycyte subtypes is well-established, the mechanism by which these cells exhibit rises in [Ca2+]i upon exposure to extracellular glucose is not completely understood (Dale,2011; Frayling et al.,2011). Recently, it was proposed that Ca2+ signaling could be mediated by hemichannels (Schalper et al.,2010) and the hemichannel activity can be modulated by the intracellular free Ca2+ concentration ([Ca2+]i) (De Vuyst et al.,2009; Schalper et al.,2008b). Hemichannels are formed by the oligomerization of six protein subunits, termed connexins (Cxs), which are a highly conserved protein family encoded by 21 genes in humans. In addition, members of a recently described three-member glycoprotein family unrelated to Cxs, termed pannexins (Panxs), can also form hemichannels in the cell membrane of vertebrates (Orellana et al.,2009).

The use of exogenous expression systems has permitted the study of electrophysiological (Sáez et al.,2005) and qualitative and quantitative permeability (Orellana et al.,2011a) properties of some hemichannels as well as mechanisms that control their functional activity. Under appropriate experimental conditions, physiologically relevant quantities of signaling molecules (e.g., ATP, glutamate, NAD+, and PGE2) are released via hemichannels to the extracellular milieu (Schalper et al.,2008a). Since hemichannels are believed to play relevant roles in intercellular signaling, the aim of the present study was to address the mechanism behind tanycyte glucosensing and determine whether it is associated with hemichannel-dependent changes in [Ca2+]i. Here, the [Ca2+]i of cultured tanycytes increased in response to glucose. This response was mediated by a transduction process involving the canonical glucosensing pathway and rapid activation of Cx43 hemichannels.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. REFERENCES
  8. Supporting Information

Reagents and Antibodies

A chemiluminescence detection kit was purchased from GE Healthcare (Aulnais-sous-Bois, France). Anti-rabbit IgG antibodies conjugated to horseradish peroxidase (HRP) was purchased from Pierce (Rockford, IL). Gap26 and 10panx1 mimetic peptides were obtained from NeoMPS, SA (Strasbourg, France). HEPES, minimal essential medium (MEM), DNAse I, poly-L-lysine, water (W3500), La3+, ethidium (Etd) bromide, thapsigargin (TG), xestospongin C (XeC), BAPTA AM, glibenclamide, xestospongin B (XeB), 4,6,-O-ethylidene-D-glucose (ETDG), alloxan, diazoxide (diazox), cytochalisin B (Cyto-B), and probenecid (Prob) were purchased from Sigma-Aldrich (St. Louis, MO). Fetal calf serum and fetal bovine serum (FBS) were obtained from Hyclone (Logan, UT). Penicillin, streptomycin, 2-deoxyglucose (2-DOG), 2-(N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)amino)-2-deoxyglucose (2-NBDG), goat anti-mouse Alexa Fluor 488 and goat anti-mouse Alexa Fluor 555, and TOPRO-3 were obtained from Invitrogen (Carlsbad, CA). 2-Deoxy-D-[1,2-(N)3H]glucose (2-[H3]DOG) was obtained from DuPont-NEN (Boston, MA). Normal goat serum was purchased from Zymed (San Francisco, CA). Cx43E2, a Cx43 hemichannel antibody (Siller-Jackson et al.,2008) was made available by Dr. Jean X. Jiang (University of Texas Health Science Center).

Animals

All animals were handled in strict accordance with the Animal Welfare Assurance (permit number 2010101A), and all animal work was approved by the appropriate Ethics and Animal Care and Use Committee of the University of Concepción, Chile. Male adult Sprague–Dawley rats were used for the experiments. Animals were kept in a 12-h light/12-h dark cycle with food and water ad libitum. Experiments in cultured cells were performed mainly at Departamento de Fisiología, Pontificia Universidad Católica de Chile (PUCC), using protocols approved by the Ethic and Biosecurity Committee of PUCC.

Cell Cultures

Cultures of hypothalamic glial cells from 1-day postnatal brains were prepared following the method described previously (García et al.,2003). After decapitation, brains were removed and the hypothalamic area was dissected to obtain a region close to the ependymal layer. The dissection was carried out from tissues immersed in dissection buffer containing 10 mM HEPES (pH = 7.4, 340 mOsm/L). Samples were incubated with 0.25% trypsin-0.2% EDTA (w/v) for 20 min at 37°C, and then it was transferred to planting medium containing MEM (Invitrogen) with 10% (v/v) FBS (Thermo Fisher Scientific, Waltham, MA) and 2 mg/mL DNAse I (Sigma-Aldrich). Cells were seeded at 1.2 × 105 cells/cm2 in culture dishes coated with 0.2 mg/mL poly-L-lysine (Sigma-Aldrich). After 2 h, the culture medium was changed to MEM supplemented with 10% FBS, 2 mM L-glutamine, 100 U/mL penicillin, 100 μg/mL streptomycin, and 2.5 μg/mL fungizone (Thermo Fisher Scientific). Mixed hypothalamic cultures were prepared using the same protocol described above; however, they were cultured in Neurobasal medium supplemented with B27 (Invitrogen). Astroglial cultures were prepared from rat brain cortex and cultured in MEM supplemented with 10% FBS. Cells were cultured in the same dish for 3 weeks, and the medium was changed every 2 days. The purity of cell cultures was evaluated using molecular markers for several cell types (Fig. 1). For all experiments, chemicals and saline solutions were prepared in ultra pure H2O.

thumbnail image

Figure 1. Immunocytochemistry characterization of cultured tanycytes. Tanycytes obtained from rat hypothalamus at 1-day postnatal were cultured for 3 weeks. (A–E) Representative confocal images depicting vimentin (A, green), GFAP (B, green), MAP2 (C, green), βIII-tubulin (D, green), Kir6.1 (E, green) in rat tanycytes under control conditions. (F) MAP2 (red) and vimentin (green) staining in mixed hypothalamic cultures of tanycytes and neurons. In blue are shown nuclei stained with TOPRO-3. (G) Quantification of immunopositive expression normalized to total cells of vimentin, Kir6.1, GLUT2, GK, MCT1, MCT4, GFAP, βIII-tubulin, MAP2, and Von Willebrand factor (VWF) in tanycytes under control conditions. Scale bar = 80 μm.

Download figure to PowerPoint

Intracellular Calcium Imaging

Cells plated on glass coverslips were loaded with 5 μM Fura-2-AM in DMEM without serum for 45 min at 37°C, and then washed three times in Locke's solution (154 mM NaCl, 5.4 mM KCl, 2.3 mM CaCl2, and 5 mM HEPES, pH = 7.4) followed by a de-esterification period of 15 min at 37°C. The experimental protocol for [Ca2+]i imaging involved data acquisition every 5 s (emission at 510 nm) at 340- and 380-nm excitation wavelengths using an Olympus BX 51W1I upright microscope with a 40× water immersion objective. Changes were monitored using an imaging system equipped with a Retga 1300I fast-cooled monochromatic digital camera (12-bit) (Qimaging, Burnaby, BC, Canada), monochromator for fluorophore excitation, and METAFLUOR software (Universal Imaging, Downingtown, PA) for image acquisition and analysis. Analysis involved determination of pixels assigned to each cell. The average pixel value allocated to each cell was obtained with excitation at each wavelength and corrected for background. Because of the low excitation intensity, no bleaching was observed even when cells were illuminated for a few minutes. The ratio was obtained after dividing the 340-nm by the 380-nm fluorescence image on a pixel-by-pixel base (R = F340nm/F380nm).

Immunofluorescence and Confocal Microscopy

Cells grown on poly-L-lysine-coated glass coverslips in 24-well plates were fixed with 4% paraformaldehyde in phosphate-buffered saline (PBS) for 30 min, washed with Tris-HCl buffer (pH = 7.8), and incubated in wash buffer containing 1% bovine serum albumin and 0.2% Triton X-100 for 5 min at room temperature. Samples were then incubated with the following primary antibodies overnight at room temperature: rabbit anti-Cx43 (1:200) and mouse anti-vimentin (1:200, DAKO, Campintene, CA), rabbit anti-GFAP (1:200, DAKO), mouse anti-MAP2 (1:50, Chemicon Temecula, CA), mouse anti-β3 tubulin (1:1,000, Promega, USA), rabbit anti-GLUT1 and GLUT2 (1:100, Alpha Diagnostic International, San Antonio, TX), chicken anti-MCT1 (1:100, Millipore, Temecula, CA), rabbit anti-MCT4 (1:20, Millipore), rabbit anti-von Willebrand factor (VWF, 1:300, Sigma) or anti-Kir1.6 (1:200, Santa Cruz Biotechnology, CA). Samples were then incubated with alexa fluor 488 or alexa fluor 555-labeled secondary antibodies and counter-stained with the DNA stain, TOPRO-3 (1:1,000, Invitrogen). Preparations were analyzed using confocal laser microscopy (D-Eclipse C1 Nikon, Tokyo, Japan).

Dye Uptake and Time-Lapse Fluorescence Imaging

For time-lapse fluorescence imaging, cells plated on glass coverslips were washed twice in PBS solution, pH = 7.4, and then bathed with Locke's solution containing 5 μM Etd. In some experiments, cells were bathed in a divalent cation-free solution (DCFS) [in (mM): NaCl (148), KCl (5), MgCl2 (1), HEPES (5), EGTA (10), pH = 7.4; at 37°C] containing 5 μM Etd. Cells on glass coverslips were placed on the microscope stage and data were acquired using the same microscope and camera described for intracellular calcium imaging. Regions of interest were placed over random cells and background was subtracted. Fluorescence was recorded every 30 s. To test for changes in slope, regression lines were fitted to points before and after various treatments using the Excel program and mean value of slopes were compared using GraphPad Prism software (GraphPad Software, San Diego, CA).

Electrophysiology

Cells platted on glass coverslips were transferred to an experimental chamber mounted on the stage of an inverted microscope (Olympus IX-51; Olympus Optical). For whole-cell voltage clamp experiments, the bath solution contained 140 mM NaCl, 5.4 mM KCl, 1 mM MgCl2, 1.8 mM CaCl2, 2 mM BaCl2, and 10 mM HEPES, pH = 7.4. The pipette solution contained 130 mM CsCl, 10 mM AspNa, 0.26 mM CaCl2, 1 mM MgCl2, 2 mM EGTA, 7 mM TEA-Cl, and 5 mM HEPES, pH = 7.2. Patch pipettes were made from borosilicate glass capillaries using a flaming/brown micropipette puller (P-87; Sutter Instruments, Union City, CA). The tip resistance was 10–15 MΩ when filled with pipette solution. Currents were filtered at 1 kHz and sampled at 5 kHz, and records were filtered with a digital low pass filter of 0.5 kHz. Data acquisition and analysis were performed using pClamp 9 (Molecular Devices, Novato, CA).

Western Blot Analysis

Cultures were rinsed twice with PBS (pH = 7.4) and harvested by scraping with a rubber policeman in ice-cold PBS containing protease and phosphatase inhibitors (1 mM orthovanadate, 10 mM α-glycerophosphate) and a complete miniprotease inhibitor (Roche Diagnostics, San Francisco, CA). Proteins were measured in aliquots of cell lysates with the Bio-Rad protein assay (Bio- Rad, Richmond, CA). Pelleted cells were resuspended in 40 μL of the protease and phosphatase inhibitor solution, placed on ice, and lysed by sonication (Ultrasonic cell disrupter, Microson, Ultrasons, Annemasse, France) after which samples were stored at −80°C or analyzed by immunoblotting. Aliquots of cell lysates or biotinylated cell surface proteins were resuspended in 1× Laemli's sample buffer, boiled for 5 min, separated on 8% SDS-PAGE and electro-transferred to nitrocellulose sheets. Nonspecific protein binding was blocked by incubation of nitrocellulose sheets in PBS-BLOTTO (5% nonfat milk in PBS). After 30 min, blots were incubated with primary antibody for 1 h at room temperature or overnight at 4°C, followed by four 15-min PBS washes. Blots were incubated with goat anti-rabbit antibody conjugated to HRP. Immunoreactivity was detected by enhanced chemiluminescence detection using the SuperSignal kit (Pierce) according to the manufacturer's instructions.

Cell Surface Biotinylation and Quantitation

Cells cultured on 90-mm dishes were washed three times with ice-cold Hank's saline solution (pH = 8.0), and 3 mL of sulfo-NHS-SS-biotin solution (0.5 mg/mL) was added followed by incubation for 30 min at 4°C. Cells were washed three times with ice-cold saline containing 15 mM glycine (pH = 8.0) to block unreacted biotin. The cells were harvested and incubated with an excess of immobilized NeutrAvidin (1 mL of NeutrAvidin per 3 mg of biotinylated protein) for 1 h at 4°C after which 1 mL of wash buffer (saline solution, pH = 7.2 containing 0.1% SDS and 1% Nonidet P-40) was added. The mixture was centrifuged for 2 min at 14,000 rpm at 4°C. The supernatant was removed and discarded, and the pellet was resuspended in 40 μL of saline solution, pH = 2.8, and containing 0.1 M glycine to release the proteins from the biotin. After the mixture was centrifuged at 14,000 rpm for 2 min at 4°C, the supernatant was collected, and the pH was adjusted immediately by adding 10 μL of 1 M Tris, pH = 7.5. Relative protein levels were measured using Western blot analysis as described above. Resulting immunoblot signals were scanned, and the densitometric analysis was performed with the Scion Image software. Densitometric arbitrary units were normalized to the signal obtained from total protein measured with Ponceau red.

Glucose Transport Assay

For 2-[H3]DOG uptake assays, cells were grown in 12-well plates to a density of 1 × 105 cells per well. Cultures were carefully selected under the microscope to ensure that only plates showing uniformly grown cells were used. Cells were washed with incubation buffer (10 mM HEPES, 135 mM NaCl, 5 mM KCl, 1.8 mM CaCl2, and 0.8 mM MgCl2) and incubated in this buffer for 30 min at room temperature. Uptake assays were performed in 500 μL of incubation buffer containing 0.2 mM 2-DOG and 3 μCi 2-[H3]DOG (30.6 Ci/mmol). Uptake was stopped by washing the cells with ice-cold PBS. Cells were lyzed in 0.5 mL of lysis buffer (10 mM Tris-HCl, pH = 8.0 and 0.2% SDS), and the incorporated radioactivity was assayed by liquid scintillation counting. Wherever appropriate, competitors and inhibitors were added to the uptake assays or preincubated with the cells.

For 2-NBDG uptake assays, cells were grown on glass coverslips. Tanycytes bathed in Locke's solution were observed with the same microscope and camera used for dye uptake experiments (see above). 2-NBDG (100 μM) was exited at 488 nm, and the emission was filtered at 505–550 nm (Porras et al., 2004). In each experiment, the resulting fluorescence was measured with Metafluor software (Universal Imaging, Downingtown, PA), and for each value, the background value was subtracted.

Measurement of Extracellular ATP Concentration

Tanycytes were grown (80% confluence) in 6-well plates for 15 days after which they were incubated in Locke's solution at 37°C for 30 min. Cells were then treated for 10 min with 10 mM glucose, and ATP concentration in the extracellular solution was measured using a luciferin/luciferase bioluminescence assay kit (Sigma-Aldrich). Baseline measurements were performed on separate cultures using standard Locke's solution. The amount of ATP in each sample was calculated from standard curves and normalized for the protein concentration using the BCA assay obtained from Pierce.

Data Analysis and Statistics

For each data group, results were expressed as mean ± standard error, and n refers to the number of independent experiments. For statistical analysis, each treatment was compared with its respective control, and significance was determined using a one-way ANOVA followed, in case of significance, by a Tukey posthoc test. For multiple group treatments, significance was determined using a two-way ANOVA followed in case of significance by a Bonferroni posthoc test.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. REFERENCES
  8. Supporting Information

Glucose Increases the Intracellular Free Ca2+ Signal in Cultured Rat Tanycytes Via GLUTs, Glucokinase, and KATP Channels

As described previously (García et al.,2003), immunohistochemical analysis revealed that cultures enriched in differentiated hypothalamic tanycytes exhibit an intense reactivity of vimentin, Kir6.1, and GLUT2, but not GFAP, MAP2, βIII-tubulin, and VWF, ruling out contamination with other hypothalamic cell types (Fig. 1A–G). In contrast, mixed hypothalamic cultures were highly immunopositive for MAP2 and vimentin (Fig. 1F), whereas cortical astroglial cultures exhibited a strong GFAP, and low vimentin reactivity (data not shown). Most tanycytes (>90%) were highly reactive to anti-GK, -MCT1, and -MCT4 antibodies (Fig. 1G), suggesting a high enrichment in β1-tanycytes, since β1- but not α-tanycytes express GK, MCT1, and MCT4 in vivo (Cortés-Campos et al.,2011; Millan et al.,2010).

Tanycytes have been proposed to mediate, at least in part, the mechanism by which the hypothalamus detects changes in extracellular glucose concentrations (Millan et al.,2010). Supporting this idea, acute in situ application of glucose increases the [Ca2+]i in α-tanycytes (Frayling et al.,2011). However, the mechanism underlying this phenomenon remains to be elucidated. Thus, we investigated whether canonical glucosensing molecules used by specialized cells, such as pancreatic β-cells, were involved in this process and if the activation of hemichannels could contribute to the glucose-induced changes in [Ca2+]i. As indicated by time-lapse measurements of Fura-2 ratio (340/380) (from now and on called Ca2+ signal), cultured tanycytes maintained in saline containing 2 mM glucose showed a low basal Fura-2 fluorescent ratio (Fig. 2A–E). However, exposure to 10 mM glucose induced a rapid, strong and transient increase in Fura-2 fluorescent ratio; peaking at ∼630% as compared with basal levels (Fig. 2A,D,E). In addition, a delay (67.6 ± 7.8 s; n = 9) between glucose addition and increase in Ca2+signal was evident (vertical gray line in Fig. 2A), suggesting that the glucose-induced rise in Ca2+signal might require activation of a signal transduction mechanism. Moreover, increasing glucose concentrations induced a proportional rise in Ca2+signal, approaching a plateau at about 20 mM glucose; peaking at ∼815% as compared with basal levels (Fig. 2D). No changes in Ca2+signal were observed upon treatment with 10 mM sucrose or mannitol ruling out the possibility of an osmolarity-induced effect (not shown).

thumbnail image

Figure 2. Glucose-induced increase in intracellular free Ca2+ signal occurs by a glucokinase- and KATP channel-dependent pathway in rat tancytes. (AC) Representative plots of relative changes in Ca2+signal (340/380 nm) over time in cultured rat tanycytes subjected to changes in glucose concentrations (horizontal bars, from 2 to 10 mM glucose) under control conditions (A), in presence of 500 μM alloxan, a glucokinase inhibitor (B) or in presence of 1 μM diazoxide (C), a KATP channel activator. In each panel, three photomicrographs of time-lapse images show changes in Fura-2 ratio (pseudo-colored scale). The delay between the addition of glucose and the increase in Fura-2 ratio is represented by the vertical gray line. (D) Averaged data normalized to control (dashed line) of maximal Fura-2 fluorescence intensity during the peak in tanycytes exposed to increasing glucose concentration. (E) Averaged data normalized to control (dashed line) of maximal Fura-2 fluorescence intensity during the peak in tanycytes exposed to 10 mM glucose alone or in combination with the following blockers: 100 μM Cyt B, 100 mM 4,6,-O-ethylidene-D-glucose (ETDG), 500 μM alloxan, 1 mM iodoacetate (IA), 200 nM AA (AA), and 1 μM diazoxide (Diazox). ***P < 0.001, effect of 10 mM glucose compared with control; #P < 0.05, ##P < 0.005, effect of blockers compared with glucose treatment. Averaged data were obtained from at least five independent experiments.

Download figure to PowerPoint

Because glucosensing by specialized cells, such as pancreatic β-cells, involves GLUT2, GK, and KATP channels (Schuit et al.,2001), the effect of inhibitors of these proteins on glucose-induced rise in Ca2+signal of tanycytes was assessed. Two GLUT inhibitors, 100 μM cytochalasin B (Cyto-B) or 100 mM ETDG (Barros et al.,2009), greatly reduced the glucose-induced increase in Ca2+signal from ∼630 to ∼234% and ∼257%, respectively (Fig. 2E). Moreover, cytochalasin E did not affect the glucose-induced rise in Ca2+signal (not shown), ruling out unspecific effects of Cyto-B on actin polymerization. These results suggest that the glucose-induced rise in Ca2+signal is triggered by glucose entry primarily via GLUTs; a small amount of glucose enters through another pathway not yet identified (see below). In addition, pretreatment with alloxan (500 μM), a GK inhibitor, completely inhibited the glucose-induced increase in Ca2+signal in tanycytes from ∼630 to ∼109% (Fig. 2B,E). To investigate if the glucose-induced rise in Ca2+signal depends on ATP generated from anaerobic glycolysis and/or oxidative phosphorylation, tanycytes were treated with iodoacetic acid (IA) or antimycin A (AA), blockers of each metabolic pathway, respectively. Pretreatment with IA (1 mM) reduced the glucose-induced rise in Ca2+signal from ∼630 to ∼107%. However, no significant reduction was observed using AA (200 nM; 612.1% ± 105.2%; Fig. 2E), suggesting that anaerobic glycolysis is the main source of ATP required for the glucose-induced rise in Ca2+signal. In pancreatic β-cells, the rise in [Ca2+]i depends on closure of KATP channels induced by an increase in intracellular ATP concentration (Schuit et al.,2001). Therefore, diazoxide (1 μM, Diazox), an activator of KATP channels, was used to examine the possible involvement of a similar mechanism in the glucose-induced rise in Ca2+signal observed in glucose-treated tanycytes. Pretreatment with Diazox for 10 min drastically reduced the glucose-induced rise in Ca2+signal from ∼630 to ∼98% (Fig. 2C,E).

Rat Tanycytes Exhibit Functional Hemichannels in Their Surface and Express Cx43 In Vitro

The rise in Ca2+ signal induced by glucose might result from opening of a Ca2+ permeable cell membrane pathway and/or Ca2+ release from intracellular stores. In glial cells the main connexin expressed is Cx43 (Giaume and Theis,2010), which has been recently shown to be Ca2+ permeable (Schalper et al.,2010). Therefore, we studied if rat tanycytes express functional Cx43 hemichannels in their surface. Similar to cortical astrocytes (Orellana et al.,2010), tanycytes cultured under control conditions exhibited a low Etd uptake rate (0.25 ± 0.02 AU/min; Fig. 3A,C). The basal Etd uptake in tanycytes under control condition was partially reduced to about 50% of the basal level by La3+ (200 μM), a general blocker of connexin hemichannels, Gap26 (200 μM), and the Cx43E2 antibody (1:500) (Fig. 3D), two Cx43 hemichannel blockers, suggesting the presence of a basal Cx43 hemichannel activity. However, the Panx1 hemichannel blockers, probenecid (Prob, 500 μM) or 10panx1 (200 μM), did not significantly alter the basal Etd uptake (97.3% ± 22.3% and 102.4% ± 6.2%, respectively), suggesting that if tanycytes express Panx1 hemichannels, they are not active under basal conditions (Fig. 3D). To address whether both Panx1 hemichannel blockers effectively inhibit Panx1 hemichannels under our conditions, we treated neurons with glutamate for 1 h, a known condition that opens Panx1 hemichannels in these cells (Orellana et al.,2011b). Both blockers inhibited the increase in the glutamate-induced Etd uptake mediated by Panx1 hemichannels (not shown).

thumbnail image

Figure 3. Rat tanycytes exhibit functional Cx43 hemichannels that are involved in the glucose-induced increase in intracellular Ca2+signal. (A,B) Fluorescence micrographs of Etd uptake (10-min exposure) in tanycytes under control conditions (A) and then exposed for 10 min to a divalent cation (Ca2+/Mg2+)-free solution (DCFS) (B). (C) Time-lapse measurement of Etd uptake in rat tanycytes under control conditions (basal) or exposed to divalent cation free solution (DCFS). La3+ (200 μM), a connexin hemichannel blocker, applied at ∼7 min of Etd uptake measurements reduced the Etd uptake to values even lower than that recorded under basal condition. (D) Averaged data normalized to control (dashed line) of Etd uptake rate by tanycytes treated with the following Cx43 hemichannel blockers co-applied during dye uptake recording: 200 μM La3+, 200 μM Gap26, and 1:500 Cx43E2; or with the following Panx1 hemichannel blockers: 200 μM 10panx1 or 500 μM probenecid (Prob). Also, it is shown the Etd uptake rate induced by DCFS conditions alone or plus the anti-Cx43 hemichannel blocker, Cx43E2. *** P < 0.001, ** P < 0.005, treatments compared with control; #P < 0.05, effect of DCFS conditions compared with blockers. Averaged data were obtained from at least five independent experiments. Scale bar = 30 μm. (E) Representative confocal image depicting vimentin (red) and Cx43 (green) immunolabeling of cultured rat tanycytes under control conditions. In blue are shown nuclei stained with TOPRO-3. Scale bar = 15 μm. (F) Representative plot of relative changes in Ca2+ signal over time in cultured tanycytes subjected to changes in glucose concentrations (horizontal bars, from 2 to 10 mM glucose) in absence of extracellular Ca2+ (Ca2+-free) (G) Averaged data normalized to control (dashed line) of maximal Fura-2 fluorescence intensity during the peak in tanycytes exposed to 10 mM glucose alone or in combination with the following blockers: 200 μM 10panx1, 500 μM probenecid (Prob), 200 μM Gap26, 1:500 Cx43E2 antibody and in absence of extracellular Ca2+. All blockers were applied 10 min before treatment with glucose. #P < 0.005, effect of blockers compared with glucose treatment. Averaged data were obtained from at least five independent experiments.

Download figure to PowerPoint

The presence of connexin hemichannels in cultured tanycytes was supported by their activation in cells exposed to a DCFS, which increases the open probability of connexin but not Panx1 hemichannels (Sáez et al.,2005). Under these conditions, a high Etd uptake rate (0.56 ± 0.01 AU/min) was found compared with control (224.3% ± 36.3%) that was almost completely blocked by Cx43E2 antibody (61.3% ± 12.1%; Fig. 3B–D), indicating that most if not all functional hemichannels are formed by Cx43.

The presence of functional Cx43 hemichannels was also in agreement with the immunofluorescence detection of Cx43 in cultured tanycytes (Fig. 3E); Cx43 was primarily localized at cell–cell interfaces, which may correspond with gap junctions (Fig. 3E). Moreover, all cultures showed vesicular Cx43 reactivity that most likely corresponds to intracellular Cx43 trafficking (Fig. 3E). Since a very small percent (∼10%) of total Cx43 is known to form hemichannels at the cell surface of cortical astrocytes (Retamal et al.,2006), it is likely that Cx43 hemichannels also correspond to a small percent of the total Cx43 protein making difficult their detection by immunofluorescence. In spite, we detected Cx43 at the cellular surface using the Cx43E2 antibody in nonpermeabilized cells (Supp. Info. Fig. 1A). In permeabilized cells, Cx43 reactivity was also detected in vesicle-like structures located at the cell interior in addition to the labeling detected at the cell periphery (Supp. Info. Fig. 1B). We also detected Cx43 hemichannels at the cellular surface using biotinylation of surface proteins and whole cell patch clamp (see below).

The Glucose-Induced Increase in Intracellular Free Ca2+ Signal Is Mediated Mainly by Ca2+ Release from Intracellular Stores and Not by Ca2+ Influx Via Hemichannels

The possible role of Cx43 hemichannels in the glucose-induced rise in Ca2+ signal was studied using specific blockers of Cx43 hemichannels, the Cx43E2 antibody (Orellana et al.,2011c; Siller-Jackson et al.,2008), and the mimetic peptide Gap26 with an amino acid sequence identical to the second loop of Cx43 (Evans et al.,2006). Cx43E2 (1:500) and Gap26 (200 μM) reduced almost completely the glucose-induced increase in Ca2+ signal in tanycytes to ∼113 and ∼107%, respectively (Fig. 3G). However, pretreatment with the preimmune antibodies (1:500) did not affect the glucose-induced increase in Ca2+ signal, supporting the specificity of Cx43E2 (not shown). In contrast, 10panx1 (200 μM) and probenecid (200 μM), both Panx1 hemichannel blockers (Pelegrin and Surprenant,2006; Silverman et al.,2008), did not reduce the glucose-induced increase in Ca2+ signal (Fig. 3G), ruling out the possible involvement of Panx1 hemichannels in the glucose-induced rise in Ca2+ signal. In spite of the strong inhibition of the glucose-induced rise in Ca2+ signal induced by Cx43E2 and Gap26, the absence of extracellular Ca2+ did not significantly reduce this response (from ∼615%, in the presence of extracellular Ca2+ to ∼590% in Ca2+-free; Fig. 3F,G), suggesting that in tanicytes exposed to glucose, Cx43 hemichannels do not allow a relevant Ca2+ inflow and thus, their role in the observed glucose-induced rise in Ca2+ signal could be indirect. In addition, the lack of effect of extracellular Ca2+-free solution on the glucose-induced rise in Ca2+ signal suggested strongly the involvement of Ca2+ release from intracellular reservoirs instead of Ca2+ influx from the extracellular medium.

Glucose Enhances the Cx43 Hemichannel Activity Via a Glucokinase/KATP Channel-Dependent Pathway in Rat Tanycytes

To elucidate the involvement of Cx43 hemichannels in the glucose-induced rise in Ca2+signal, the effect of glucose on the hemichannel activity was assessed. Under control conditions (2 mM glucose), tanycytes exhibited low Etd uptake (Fig. 4A,D), which prominently increased upon treatment with 10 mM glucose (from 0.25 ± 0.02 to 0.57 ± 0.02 AU/min, n = 6) (Fig. 4B,D,F). An important difference with the glucose-induced rise in Ca2+ signal, which showed a delay of about 1 min (Fig. 4A,B), was that the delay of the glucose-induced rise in Etd uptake was <20 s (Fig. 4D). Moreover, as in the case of the Ca2+ signal changes, the Etd uptake increase was glucose concentration-dependent, approaching a plateau at 40 mM glucose (425.4% ± 182.3%, compared with basal; Fig. 4E). In these cells, no concurrent changes in intercellular communication mediated by gap junctions, evaluated by the intercellular transfer of microinjected Lucifer yellow, was observed (Supp. Info. Fig. 2A–D).

thumbnail image

Figure 4. Increased Etd uptake induced by glucose is mediated by a glucokinase/KATP channel-dependent pathway in cultured rat tanycytes. (AC) Representative immunofluorescence images depicting vimentin (white) labeling and Edt (red) nuclei-staining from dye uptake experiments (10-min exposure to dye) in cultures of tanycytes under control conditions (A) or treated with 10 mM glucose for 5 min alone (B) or glucose plus 1:500 Cx43E2, an anti-Cx43 antibody that blocks Cx43 hemichannel (C). (D) Time-lapse measurement of Etd uptake in rat tanycytes treated with 2, 10, or 20 mM glucose. La3+ (200 μM) applied at ∼10 min of Etd uptake measurement inhibited dye uptake. (E) Average of Etd uptake rate normalized to control (dashed line) in tanycytes exposed to increasing concentrations of glucose. * P < 0.001, treatments compared with control. (F) Averaged Etd uptake rate normalized to control (dashed line) by tanycytes treated with 10 mM glucose alone or in combination with the following blockers: 100 μM cytochalasin B (Cyt B), 100 mM ETDG, 500 μM alloxan, 1 mM iodocetic acid (IA), 200 nM antimycin A (AA), 1 μM diazoxide (Diazox), 200 μM La3+, 200 μM Gap26, 1:500 Cx43E2, 200 μM 10panx1, or 500 μM Prob. ** P < 0.005, 10 mM glucose treatment compared with control; #P < 0.05, ##P < 0.005, effect of 10 mM glucose treatment compared with blockers. Averaged data were obtained from at least four independent experiments. Scale bar = 15 μm.

Download figure to PowerPoint

To evaluate if the glucosensing proteins, GLUTs, GK, and KATP channels, are involved in the glucose-induced Etd uptake, specific drugs against each of these molecular components of the glucosensor transduction system were employed. As compared with control, 100 μM Cyto-B, 100 mM ETDG, 500 μM alloxan, 1 mM IA, and 1 μM Diazox reduced the glucose-induced Etd uptake from ∼211 to ∼138, ∼141, ∼103, ∼105, and ∼109%, respectively (Fig. 4F). However, 200 nM AA did not affect it (207.2% ± 19.4%, n = 4; Fig. 4F). Therefore, the glucose-induced Etd uptake requires the functional participation of GLUTs, GK, and glycolysis-derived ATP to block KATP channels. Relevant to this interpretation is that inhibitors of these glucose-sensing proteins did not alter the DCFS-induced Etd uptake (Supp. Info. Fig. 3), indicating that they act as glucosensing modulators and not as hemichannel blockers.

Glucose Increases the Macroscopic Cx43 Hemichannel Current, but Does Not Alter the Unitary Conductance or the Surface Cx43 Levels in Tanycytes

To identify the pathway through which glucose increases the membrane permeability of tanycytes, whole-cell voltage-clamp studies were undertaken, and the macroscopic membrane current (total current measured in the absence of other active membrane channels) and the presence of hemichannel unitary events applying voltage ramp protocols from −80 to +80 mV were assessed (Fig. 5A).

thumbnail image

Figure 5. Glucose increases Cx43 hemichannel currents, but does not affect the surface Cx43 levels in cultured rat tanycytes. (A) Voltage ramps from −80 to +80 mV, 5 s in duration, were applied. Each ramp was initiated and finished by a transition from 0 to −80 and +80 to 0 mV, respectively. Membrane current was measured in a whole-cell voltage-clamp configuration using low-density cultures of rat tanycytes. (B) Voltage ramp from −80 to +80 mV, 5 s duration in tanycytes under control conditions (black line) or treated for 3 min with 10 mM glucose alone (magenta line) or 10 mM glucose plus 1:500 Cx43E2 (green line). A record section at about +60 mV the current trace was point-by-point converted in conductance values showing unitary events of about 220 pS. (C) Cultured tanycytes were treated for 5 min with 10 mM glucose. Total (left panel) and surface (right panel) levels of Cx43 in tanycytes under control conditions (lane 1) or treated for 5 min with 10 mM glucose alone (lane 2), 10 mM glucose plus 500 μM alloxan (lane 3) or plus 1 μM diazoxide (Diazox) (lane 4). The Cx43 phosphorylated (P1-P2) and nonphosphorylated (NP) forms are indicated in the left. Total levels of each analyzed protein were normalized according to the levels of α-tubulin detected in each lane. Surface levels of each analyzed protein were normalized according to the total protein loaded as revealed by staining with Ponceau red (PR) in each lane. Similar observations were made in two other independent experiments. (D) Averaged data on current events at +60 (white bars) or −60 mV (black bars) in tanycytes under control conditions, exposed to 10 mM glucose alone or 10 mM glucose plus Cx43E2. * P < 0.001, glucose treatment compared with control; #P < 0.001, inhibitory effect of blocker on the glucose-induced response. Averaged data were obtained from three independent experiments in which at least eight cells were analyzed per treatment. (E) Quantification of total (white bars) and surface (black bars) levels of Cx43 normalized to control (dashed line) in tanycytes treated for 5 min with 10 mM glucose alone, plus 500 μM alloxan or 1 μM Diazox. Averaged data were obtained from three independent experiments.

Download figure to PowerPoint

In control tanycytes, unitary current events were not detected in 17 cells (not shown); however, six cells showed a few small unitary transitions at negative potentials (Fig. 5B). In contrast, in tanycytes treated with 10 mM glucose for 3 min the total current was much bigger than in control cells, and numerous unitary current events were detected at either negative (at −60 mV from ∼15 to ∼197 pA, n = 12 cells; Fig. 5B,D) or positive potentials (at +60 mV from ∼14 to ∼205 pA, n = 12 cells; Fig. 5B,D). In most cells (n = 18), unitary current events were recorded, and point-by-point conversion of current to conductance values revealed single channels of 218 ± 6 pS (see inset in Fig. 5B). This value is close to the expected unitary conductance (∼220 pS) of Cx43 hemichannels (Contreras et al.,2003). To elucidate the involvement of such hemichannels in the glucose-induced currents, tanycytes were treated acutely with Cx43E2 (1:500). This Cx43 hemichannel blocker reduced almost completely the rise in macroscopic current (at −60 mV from ∼197 to ∼27 pA and at +60 mV from ∼205 to ∼25 pA, n = 12) and the appearance of discrete unitary current transitions induced by glucose (Fig. 5B,D), indicating that glucose induces activation of Cx43 hemichannels. Importantly, the reversal potential of all the above mentioned currents was close to 0 mV. However, the glucose-induced currents presented a reversal potential close +5 mV (Fig. 5B). Moreover, 10 mM glucose increased the frequency of unitary currents of Cx43 hemichannels with conductance around ∼220 pS (Supp. Info. Fig. 4A–D), whereas did not significantly alter the decay constant (τ) of open time compared with control (3.7 ± 0.7 and 5.2 ± 0.6 ms, respectively; Supp. Info. Fig. 4E,F). Furthermore, 10 mM glucose increased the mean open time compared with controls conditions (0.7 ± 0.1 and 0.20 ± 0.03 ms, respectively).

The increase in Cx43 hemichannel activity might be due to an increase in open probability per hemichannel or/and increase in the number of hemichannels present in the cell membrane. Previous studies have associated hemichannel-mediated dye uptake with increased surface levels of hemichannels (Orellana et al.,2010). Therefore, the effect of glucose on the total and surface levels of Cx43 in tanycytes was evaluated. Neither total nor surface levels of Cx43 were affected by treatment with 10 mM glucose for 5 min (Fig. 5C,E, respectively, n = 3). Also, neither 500 μM alloxan nor 1 μM diazox in combination with glucose affected total Cx43 or surface levels in tanycytes (Fig. 5C,E, respectively, n = 3). Importantly, 10 mM glucose might decrease the total nonphosphorylated Cx43 and increase the total phosphorylated form of Cx43 (Fig. 5C); however, total levels of Cx43 were not significantly affected (Fig. 5E).

Tanycytes Are Permeable to Glucose Via GLUTs and Cx43 Hemichannels

Since neither the rise in Ca2+ signal nor the increase in Etd uptake induced by glucose were completely inhibited with GLUT blockers (Figs. 2F and 4F) and because other glial cells also present glucose uptake through Cx43 hemichannels (Retamal et al.2007), the role of these channels as an alternative pathway for glucose entry into tanycytes was explored. Specifically, uptake of 2-DOG and 2-NBDG in tanycytes under control conditions or bathed in DCFS to enhance the open probability of Cx43 hemichannels (Sáez et al.,2005) was measured. After 1 min, 2-DOG uptake close to 20–30 nmoles per million of cells (22.1 ± 8 nmol/106 cells; Fig. 6A) and low 2-NBDG uptake (37.4 ± 7.3 AU, n = 6) was observed under control conditions (Fig. 6B). Uptake of 2-DOG and 2-NBDG was drastically reduced by 100 mM ETDG (∼6nmol/106 cells and ∼9 AU, respectively, n = 5) or 100 μM Cyto-B (∼7 nmol/106 cells and ∼7 AU, respectively, n = 5), whereas Cx43E2 (1:500) only partially inhibited 2-DOG and 2-NBDG uptake (∼15 nmol/106 cells and ∼22 AU, respectively, n = 3; Fig. 6). Notably, co-application of Cyto-B and Cx43E2 almost completely blocked the uptake of 2-DOG and 2-NBDG (∼0.8 nmol/106 cells and 0.8 AU, respectively, n = 3; Fig. 6B).

thumbnail image

Figure 6. Glucose uptake occurs via GLUTs and Cx43 hemichannels in rat tanycytes. Averaged data at 1 min of 2-DOG (A) and 2-NBDG (B) uptake in tanycytes under control conditions or treated with 100 mM ETDG, 100 μM cytochalasin B (Cyto-B), 1:500 Cx43E2 or 1:500 Cx43E2 plus 100 μM Cyto-B. In addition, 2-DOG and 2-NBDG uptake by tanycytes exposed to DCFS conditions for 1 min alone or plus 100 mM (ETDG), 100 μM Cyto-B, 1:500 Cx43E2 or 1:500 Cx43E2 plus 100 μM Cyto-B was analyzed. *** P < 0.001, ** P < 0.005, * P < 0.05; treatments compared with control; #P < 0.05, treatments compared with DCFS conditions. Averaged data were obtained from at least four independent experiments.

Download figure to PowerPoint

As predicted, tanycytes bathed in DCFS exhibited higher 2-DOG (∼38 nmol/106 cells, n = 5) and 2-NBDG (∼66 AU, n = 5) uptake than tanycytes under control conditions (Fig. 6A,B). Under these conditions, uptake of 2-DOG or 2-NBDG was more prominently inhibited by Cx43E2 (∼14 nmol/106 cells and ∼24 AU, respectively, n = 3), than by ETDG (∼22 nmol/106 cells and ∼31 AU, respectively, n = 5) or Cyto-B (∼24 nmol/106 cells and ∼27 AU, respectively, n = 5; Fig. 6). Furthermore, in cells treated with Cyto-B plus Cx43E2 the uptake of 2-DOG and 2-NBDG was almost completely blocked (∼1 nmol/106 cells and ∼0.7 AU, respectively, n = 3).

The Glucose-Induced Increase in Intracellular Free Ca2+ Signal Occurs Via ATP Released Through Cx43 Hemichannels

Because in parallel experiments the onset of glucose-induced Etd uptake occurred before (Fig. 4D) than the rise in Ca2+ signal (Fig. 2A), simultaneous measurements of Etd uptake and Ca2+signal in tanycytes exposed to glucose were performed to better elucidate the onset of these sequential changes. In these studies, the Etd uptake induced by 10 mM glucose occurred at ∼12 s (Fig. 7A–H,I, n = 8), whereas the glucose-induced rise in Ca2+ signal was observed at ∼67 s (Fig. 7A–H,I, n = 9). Because the glucose-induced changes did not depend on the extracellular [Ca2+] (Fig. 2B,F), the above results suggest that opening of Cx43 hemichannels results in the extracellular release of a molecule that reaches a concentration sufficient to trigger Ca2+ release from intracellular stores.

thumbnail image

Figure 7. Glucose increases the intracellular free Ca2+ signal in tanycytes by ATP released via Cx43 hemichannels. (AH) Representative fluorescence micrographs of simultaneous time-lapse imaging showing changes in Fura-2 ratio (A–D, pseudo-colored scale) and Etd uptake (E–H, red) in tanycytes treated with 10 mM glucose for 0 (A and E), 48 (B and F), 50 (C and G), and 200 s (D and H). (I) Simultaneous plots of relative changes in [Ca2+]i and Etd uptake over time of cells 1 (red), 2 (green), 3 (yellow), and 4 (magenta) depicted in panel A. The delay between the increase in Etd uptake and Fura-2 ratio is indicated by the vertical dashed line. (J) Averaged data normalized to control (dashed line) of maximal Fura-2 fluorescence intensity during the peak in tanycytes exposed to 10 mM glucose alone or in combination with the followed blockers: 10 U/mL apyrase (Apyr), 200 μM oxidized ATP (oATP), 10 μM (BBG), 200 μM suramin (Sur), 10 μM MRS2179, 2 μM thapsigargin (TG), 5 μM xestospongin C (XeC), and 5 μM xestospongin (XeB). ** P < 0.005, 10 mM glucose treatment compared with blockers. (K) Averaged values of ATP released by tanycytes under control conditions or treated with 10 mM glucose alone or in combination with the following blockers: 100 μM cytochalisin B (Cyt B), 500 μM alloxan, 1 mM IA, 1 μM diazoxide (Diazox), 200 μM 10panx1, 500 μM probenecid (Prob), 200 μM La3+, 200 μM Gap26 or 1:500 Cx43E2. *** P < 0.001, 10 mM glucose treatment compared with control; ##P < 0.005, effect of 10 mM glucose treatment compared with blockers. Averaged data were obtained from at least four independent experiments. Scale bar = 35 μm.

Download figure to PowerPoint

Previous reports have demonstrated that Cx43 hemichannels can mediate the release of ATP (Kang et al.,2008). Because ATP can induce Ca2+ release from intracellular stores via P2Y receptors, ATP release by tanycytes in response to glucose was examined. The increase in Ca2+ signal induced by glucose was prevented by 10 U/mL apyrase, a phosphatase that degrades ATP, and 200 μM suramine, an inhibitor of both P2Y and P2X receptors, peaking at ∼135 or ∼126%, respectively (Fig. 7J). However, no differences were observed using 200 μM oATP, a general P2X receptor blocker, or 10 μM brilliant blue G (BBG), a P2X7 receptor blocker (∼610 or ∼624%, respectively; Fig. 7J). These results indicated that ATP and P2Y receptors are crucial for glucose-induced increases in Ca2+ signal. In support of this interpretation, MRS2179, a P2Y1 receptor blocker at concentrations up to 30 μM, inhibited the glucose-induced rise in Ca2+ signal (129.2% ± 17.3%; Fig. 7J). Moreover, 2 μM thapsigargin, a compound that depletes the intracellular Ca2+ stores, or 5 μM xestospongin C (XeC) and 5 μM xestospongin B (XeB), both IP3 receptor blockers, inhibited the rise in Ca2+ signal induced by glucose (∼158, ∼125, and ∼120%, respectively; Fig. 7J), suggesting the involvement of IP3-mediated Ca2+ release.

Then, the effect of glucose on ATP release via Cx43 hemichannels and glucosensing signaling was examined. The extracellular medium of tanycytes treated with 10 mM glucose exhibited higher ATP levels than that of control tanycytes (from ∼1 pmol/106 cells to ∼45 pmol/×106 cells, n = 6; Fig. 7K). The increase in extracellular ATP concentration induced by glucose was prevented by 100 μM Cyto-B, 500 μM alloxan, 1 mM IA, 1 μM Diazox, and Cx43E2 (∼7, ∼1, ∼0.9, ∼0.2, and ∼0.2 pmol/106, respectively, n = 6; Fig. 7K).

The effect of ATP itself on the Ca2+ signal in tanycytes was also examined; 100 μM ATP induced a rise in Ca2+ signal similar to that observed with 10 mM glucose (∼682%, n = 3; Fig. 8A,B). Moreover, the ATP-induced rise in Ca2+ signal was inhibited by 200 μM suramine (∼133%, n = 3), 2 μM thapsigargin (∼147%, n = 3), 5 μM XeC (∼156%, n = 3), and 5 μM XeB (∼130%, n = 3), but not by 200 μM oATP or 10 μM BBG (∼671 and ∼656%, respectively, n = 3; Fig. 8B). Thus, the results support the hypothesis that proposes extracellular ATP as mediator of the glucose-induced rise in [Ca2+]i.

thumbnail image

Figure 8. ATP increases the intracellular free Ca2+ signal in tanycytes via P2Y receptors. (A) Representative plots of relative changes in Ca2+ signal over time in cultured rat tanycytes subjected to 100 μM ATP under control conditions or pretreated with MRS21. (B) Averaged data normalized to control (dashed line) of maximal Fura-2 fluorescence intensity during the peak in tanycytes exposed to 100-μM ATP alone or in combination with the followed blockers: 200 μM oxidized ATP (oATP), 10 μM brilliant blue G (BBG), 200 μM suramine (Sur), 10 μM MRS2179, 2 μM thapsigargin (TG), 5 μM xestospongin (XeC), and 5 μM xestospongin (XeB). * P < 0.001, 100 μM ATP treatment compared with control; #P < 0.001, 100 μM ATP treatment compared with blockers.

Download figure to PowerPoint

Because of the complexity behind the activation of Cx43 hemichannels of this system, we investigated whether an IP3 receptor-mediated Ca2+ signal is needed to promote Cx43 hemichannel opening. XeC and XeB did not modify the glucose-induced hemichannel activity (Supp. Info. Fig. 5). Similar results were obtained in tanycytes loaded with BAPTA (Supp. Info. Fig. 5), indicating that cytoplasmic Ca2+ does not mediate the activation of Cx43 hemichannel opening. In contrast, as it was shown before, KATP channel inhibition was necessary for the glucose-induced hemichannel activity. However, glibenclamide, an inhibitor of KATP channels, by itself did not induce hemichannel opening (Supp. Info. Fig. 5), suggesting that an additional mechanism activated almost simultaneously, and not yet identified, is necessary for Cx43 hemichannel activation.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. REFERENCES
  8. Supporting Information

The present study demonstrates that tanycytes possess a rapid, glucose-activated signal transduction pathway, which includes GLUTs, GK, anaerobic glycolysis-derived ATP, KATP channels, Cx43 hemichannels, extracellular ATP, P2Y receptors, and IP3 activated intracellular Ca2+ stores.

Recently, glucose and nonmetabolizable analogs of glucose were demonstrated to increase the [Ca2+]i in hypothalamic slices, specifically in α tanycytes located in the most dorsal region of the hypothalamus (Frayling et al.,2011), suggesting that the pancreatic β-cell paradigm does not apply to these cells. However, in the present study, glucose increased the [Ca2+]i in a concentration-dependent manner in cultured β-tanycytes by a process in which glucose uptake mainly via GLUT is critical as Cyto-B and ETDG strongly inhibited this response. This is in agreement with the reported expression of GLUT1 and the glucosensing transporter, GLUT2, by β1-tanycytes (García et al.,2003). Because the glucose-induced rise in Ca2+ signal was prevented by a GK blocker as well as a KATP channel activator, these two proteins may play an essential role in tanycyte glucosensing, which is also consistent with the high in vivo expression of these proteins by these cells (Marty et al.,2007; Thomzig et al.,2001). In addition, in vivo pharmacological and molecular inhibition of GK or KATP channels impairs feeding behavior in rodents (Miki et al.,2001; Sanders et al.,2004). The difference between the results of the present study and those of Frayling et al. (2011) could be explained by the type of tanycyte population used and the model (ex vivo versus in vitro). Frayling et al. (2011) used α-tanycytes from brain slices, whereas our cultures were highly enriched in β1-tanycytes, which express the same enzymes and transporters detected in situ (e.g., GLUT1, GLUT2, MCT1, MCT4, GK; Cortés-Campos et al.,2011; García et al.2001, 2003; Millan et al.,2010). In this regard, tanycytes may sense glucose by more than one mechanism, which is determined by the predominant type of tanycyte.

In the present study, glucose-induced Etd uptake was not completely inhibited when GLUTs were blocked, suggesting the contribution of another source of glucose transport/diffusion through the cell membrane. Accordingly, the glucose uptake was completely inhibited when both GLUTs and connexin hemichannels were blocked, suggesting that under control conditions, Cx43 hemichannels participate in glucose diffusion toward the cytoplasm of tanycytes to a minor extent despite the increase in Cx43 hemichannel activity, while GLUTs are the main protagonist in this transport.

Intracellular ATP generated by anaerobic glycolysis but not produced by oxidative phosphorylation was required for the glucose-induced rise in Ca2+signal since the effect of glucose was prevented by IA and not by AA. This finding is similar to that reported for pancreatic β-cells (Mertz et al.,1996). The high glycolytic activity of tanycytes has been previously demonstrated; they release lactate at physiological concentrations of glucose, which is inhibited by classical MCT inhibitors (Cortés-Campos et al.,2011). In pancreatic β-cells, the Ca2+ influx proceeds through voltage-dependent Ca2+ channels and is critical for glucosensing (Schuit et al.,2001). However, data from the present study indicate that extracellular Ca2+ influx is not involved in this response in tanycytes, since the glucose-induced rise in Ca2+ signal was observed even in the absence of extracellular Ca2+, indicating the involvement of intracellular Ca2+ stores. Nevertheless, the glucose-induced rise in Ca2+ signal was not detected after Cx43 hemichannel blockade, implying that they are involved in changing the [Ca2+]i by allowing release of a paracrine factor likely to be ATP. In agreement with this interpretation, an increase in Cx43 hemichannel activity was observed in tanycytes treated with glucose and the increase in Ca2+ signal and Etd uptake were completely inhibited by several Cx43 hemichannel blockers (e.g., La3+, Gap26, and Cx43E2). In contrast, Panx1 hemichannel blockers were without effect, indicating for the first time that tanycytes exhibit functional Cx43 hemichannels, which is consistent with the high Cx43 immunoreactivity detected in these cells. In support of the idea that Cx43 hemichannels are involved in glucosensing, patch clamp experiments revealed an increase in macroscopic membrane current induced by glucose. This is most likely due to opening of Cx43 hemichannels already present in the cell surface due to the following: (1) it was completely inhibited by Cx43E2, a specific blocker of Cx43 hemichannels (Orellana et al.,2011c; Siller-Jackson et al.,2008); (2) the total surface levels of Cx43 hemichannels were not affected by glucose; and (3) the unitary events were of ∼ 220 pS, characteristics that correspond to that of Cx43 hemichannel (Contreras et al.,2003). Thus, it is likely that the glucose-induced increase in total current was due to an increase in open probability of each Cx43 hemichannel already present at the cell surface and not to recruitment of more hemichannels at the cell surface or increase in unitary hemichannel conductance. Unlike cortical astrocytes that express hemichannels formed by Panx1 (Iglesias et al., 2009; Iwabuchi and Kawahara, 2011) or Cx43 (Orellana etal.,2011c; Retamal et al.,2007), cultured tanycytes seems to express mainly hemichannels formed by Cx43. Theslightly positive reversal potential observed in tanicytes treated with glucose might reflect either a change Cx43 hemichannel selectivity or the presence of a more selective membrane channel simultaneously activated by glucose and co-recorded with Cx43 hemichannels.

Since the glucose-induced hemichannel activity was inhibited by blockers of the canonical glucosensing pathway (e.g., GLUTs, GK, and KATP channels), activation of Cx43 hemichannels likely occurs downstream of the glucosensing molecules (Fig. 9). A puzzling aspect is how KATP channel inhibition leads to increased Cx43 hemichannel activity. One possibility is that it might result from membrane depolarization as consequence of KATP channel inhibition by intracellular ATP. In agreement with this possibility, Cx43 hemichannels expressed in HeLa cells are activated by positive potential (Contreras et al.,2003). However, in glucose-treated tanycytes, the activity of Cx43 hemichannels was also high at negative potentials, suggesting that membrane depolarization is not the only possible mechanism involved in enhancing Cx43 hemichannel activity. In support to this notion, membrane depolarization under control conditions did not enhance the membrane current mediated by Cx43 hemichannels. Moreover, glibenclamide by itself did not induce hemichannel opening, suggesting that both glucose or its derivates (e.g., ATP) and inhibition of KATP channels are necessary for Cx43 hemichannel activation. Among the putative mechanism, it is possible that glucose or its derivates activate kinases/phosphatases regulating the phosphorylation state of Cx43 hemichannels, which are known to affect the gating mechanism of these channels (Sáez et al.,2010). However, further studies are required to examine this and other alternative possibilities to identify the mechanism by which Cx43 hemichannels are activated by glucose in tanycytes.

thumbnail image

Figure 9. Model of glucosensing mechanism in tanycytes. Upon a rise in extracellular glucose concentration, glucose enters (1) mainly through GLUTs and to a minor extent via Cx43 hemichannels, leading to glucokinase-dependent phosphorylation of glucose (2). The ATP generated during glycolysis (3) via processing of glucose-derived substrates causes closure of KATP channels (4). This event promotes by an unknown mechanism the opening of Cx43 hemichannels (5). The ATP released via Cx43 hemichannels (6) activates P2Y receptors (7) and induces formation of cytoplasmic inositol (1,4,5)-trisphosphate (IP3), which promotes the release of Ca2+ stored in the endoplasmic reticulum (8), raising the [Ca2+]i (9).

Download figure to PowerPoint

Acute glucose increases the release of ATP in tanycytes (Frayling et al.,2011). Accordingly, the present study shows that the extracellular ATP concentration is enhanced in tanycyte cultures treated with glucose and most likely is the consequence of more ATP released via Cx43 hemichannels, because inhibitors of Cx43 hemichannels and not of Panx1 hemichannels prevented the glucose-induced increase in extracellular ATP levels. This increase in extracellular ATP concentration was also prevented by MRS2179, thapsigargin and xestospongin C, revealing the involvement of the metabotropic P2Y1 receptors, IP3 receptors and intracellular Ca2+ stores, respectively. Supporting the possible regulatory role of ATP in feeding behavior, intracerebroventricular infusion of this molecule stimulates neurons localized in the lateral hypothalamic area (Wollmann et al.,2005), dorsomedial hypothalamic nucleus (Matsumoto etal.,2004), and VMN (Sorimachi et al.,2001). Moreover, extracellular ATP and possibly ADP may regulate food intake via activation of P2Y1 receptors (Kittner et al.,2006). ATP release by tanycytes may modulate neuronal activity in hypothalamic areas associated with feeding behavior, including the AN and VMN, which are in close contact with these cells. If so, the activation of purinergic P2Y receptors might be turned off in part by diffusion of ATP/ADP to distal regions as well as by desensitization of P2Y1 receptors (Choi et al.,2008) and degradation of extracellular ATP/ADP by extracellular phosphatases in the extracellular compartment of tanycytes (Firth and Bock,1976).

Future in vivo studies will be required to determine whether tanycytes could sense extracellular changes in glucose concentration and transmit them to neurons via Ca2+ waves and/or the release of paracrine factors (e.g., ATP). The present work shows the involvement of several glucosensing proteins in the glucose-induced rise in [Ca2+]i on β1-tanycytes (Fig. 9), which might contribute to the knowledge on the glucosensing mechanism in the brain and may open novel pharmacological strategies for therapeutic treatment of feeding behavior disorders.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. REFERENCES
  8. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. REFERENCES
  8. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
GLIA_21246_sm_SuppFig1.tif3133KSupporting Information Figure 1. C×43E2 antibody detects C×43 in rat tanycytes. Representative confocal image depicting C×43 (green) and vimentin (red) immunolabeling of permeabilized (A) or nonpermeabilized (B) tanycytes using the C×43E2 antibody. Scale bar = 20 μm.
GLIA_21246_sm_SuppFig2.tif2938KSupporting Information Figure 2. Glucose does not alter gap junction communication in rat tanycytes. The functional state of gap junction channels was examined by microinyectin of Lucifer yellow (LY) in cultured tanycytes. Fluorescence (A) and phase contrast (B) micrographs of LY coupling in tanycytes under control conditions. * in panel A denotes the cell microinjected with LY. Scale bar = 40 μm. (C) Coupling index (white bar, mean number of cells to which LY spread) and incidence of coupling (black bar, percentage of injections that resulted in LY transfer) by tanycytes under control conditions. (D) Coupling index (white bar) and incidence of coupling (black bar) normalized to control (dashed line) by tanycytes subjected to increased glucose concentrations. Averaged data were obtained from three independent experiments, in which LY was microinjected into at least 10 cells.
GLIA_21246_sm_SuppFig3.tif1016KSupporting Information Figure 3. Blockers of GLUTs and glucosensing pathway do not inhibit tanycyte hemichannels. Averaged data normalized to control (dashed line) of Etd uptake rate by tanycytes subjected to DCFS conditions for 10 min alone or in combination with the following blockers: 100 μM cytochalisin B (Cyto-B), 100 mM ETDG, 500 μM alloxan, 1 mM iodocetic acid (IA) and 1 μM diazoxide (Diazox). Averaged data were obtained from three independent experiments.
GLIA_21246_sm_SuppFig4.tif714KSupporting Information Figure 4. Records of unitary Cx43 hemichannel currents in tanycytes exposed to glucose. Whole-cell voltage clamp recordings after the application of a voltage step from 0 to +80 mV a tanycyte under control conditions (A) or treated for 3 min with 10 mM glucose (B). C and D show the frequency (counts) distribution of conductance values of records shown in A and B, respectively. The conductance values of fully open state corresponds to ∼220 pS. (E and F) Open time distributions for a C×43 hemichannel at holding potential of +80 mV in tanycytes under control conditions (E) or treated for 3 min with 10 mM glucose (F) . The decay constant (t) for open time was calculated by exponential regression with one constant. The numbers indicate the mean decay constants (τ).
GLIA_21246_sm_SuppFig5.tif535KSupporting Information Figure 5. IP3-receptor inhibitors do not inhibit the glucose-induced hemichannel activity. Averaged data normalized to control (dashed line) of Etd uptake rate in tanycytes treated with 10 mM glucose alone or in combination with the following blockers: 10 μM BAPTA AM, 5 μM xestospongin (XeC) and 5 μM xestospongin (XeB). The application of 10 μM glibenclamide (Glib) did not affect the Etd uptake observed in tanycytes under control condition. Averaged data were obtained from three independent experiments.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.