Insulin Regulates Astrocytic Glucose Handling Through Cooperation With IGF-I

Contrary to its glucoregulatory actions in peripheral tissues, physiological variations of circulating insulin levels do not modulate glucose handling by brain cells (1). Only under pathological circumstances, such as diabetes, do brain glucoregulatory actions of insulin manifest (2). Therefore, mechanisms of neurovascular coupling, whereby active brain regions locally increase glucose and oxygen uptake, are not considered to involve insulin under physiological circumstances. Moreover, both endothelial cells and astrocytes, the main cellular constituents of the blood-brain barrier (BBB) express GLUT1 as their main facilitative transporter (3), and GLUT1 is considered to be largely insulin insensitive (4). At the same time, there is evidence that the structurally related peptide IGF-I affects glucose metabolism in the brain (5). Importantly, in animal models of diabetes, not only is there brain insulin resistance (6), but also brain IGF-I levels are reduced (7). We now describe a cooperative mechanism whereby insulin stimulates glucose uptake by forebrain astrocytes acting in concert with IGF-I. Cooperation of insulin and IGF-I (I+I) may explain brain glucose uptake on demand without changes in circulating insulin levels.

Adult (3–5 months old) and newborn C57BL6/J mice were used. To obtain mice with glial fibrillary acidic protein (GFAP)–specific deletion of PKD1, PKD1^loxP/loxP mice (8) were crossed with PKD1^loxP/loxP/GFAP-Cre mice. GFAP-Cre mice where obtained from The Jackson Laboratory (B6.Cg-Tg(Gfap-cre)73.12Mvs/J: stock #012886). Only male PKD1^loxP/loxP (PKD1^floxed) and PKD1^loxP/loxP/GFAP-Cre (PKD1^Δ) mice were used. Mice were of 129SvEv/C57BL/6 mixed background. Genotyping and recombination analysis in the PKD1^loxP/loxP/GFAP-Cre mouse brain and in primary cultures of astrocytes was performed by PCR using specific pairs of primers (8). As a positive control for genomic amplification, oIMR7339 primers were used (The Jackson Laboratory protocols). Two-month-old male mice from both genotypes were used for brain cortex dissection and preparation of DNA and protein extracts, as well as for positron emission tomography (PET) analysis. Primary cultures of astrocytes were obtained from brain cortices of P₁-P₂ pups bearing the two genotypes. Animal procedures followed European Council Directives (86/609/EEC and 2003/65/EC) and approval of the local bioethics committees.

Plasmids transfected into astrocytes were as follows: IGF-I KR, a negative-dominant form of the IGF-I receptor (IGF-IR) previously described (9). HIR-K1030R4 (10), a negative-dominant form of the IGF-IR, was acquired from Addgene (Cambridge, MA), and GLUT1-Exo Flag was a gift from Jeffrey C. Rathmell (11). Short hairpin RNAs (shRNAs) against GLUT1 were from Origene (Rockville, MD); the shRNA against GAIP-interacting protein C terminus (GIPC) was constructed (as described at www.addgene.org/tools/protocols/plko/) using the following primers: 5′CCGGACTCACCG AACCTCGGAAGGCCTCGAGGCCTTCCGAGGTTCGGTGAGTTTTTTG3′ and 5′AATTCAAAAAACTCACCGAACCTCGGAAGGCCTCGAGGCCTTCCGAGGTTCGGTGAGT3′ directed against the 684–704 fragment of GIPC mRNA.

Astroglial cultures with >95% GFAP-positive cells were prepared as described previously (9). Postnatal (day 3–4) brains were dissected, and the cortex and hippocampus were removed and mechanically dissociated. The resulting cell suspension was centrifuged and plated in DMEM/F-12 (Life Technologies) with 10% FBS (Life Technologies) and a 100 mg/mL antibiotic-antimycotic solution (Sigma-Aldrich, Madrid, Spain). After 15–20 days, astrocytes were replated at 1.2 × 10⁵ cells/well. Pure cerebellar granule neurons from postnatal mouse cerebellum (>99% β3-tubulin–positive cells) were obtained as described previously (12). Briefly, P7 brains were dissected, and cerebella were removed, dissociated mechanically, and then enzymatically digested with papain solution (Worthington Biochemical Corporation, Lakewood, NJ). Finally, neurons were plated in wells covered with poly-l-lysine in medium with Neurobasal plus B27 medium (Life Technologies), glutamine, and 25 mmol/L KCl. Endothelial cell cultures were performed as described previously (13). Briefly, dissection was performed on ice, and cortices were cut into small pieces (1 mm³) and digested in a mixture of collagenase/dispase (270 units collagenase/mL and 0.1% dispase) and DNAse (10 units/mL) in DMEM (Life Technologies) for 1.5 h at 37°C. The cell pellet was separated by centrifugation in 20% BSA/DMEM (1,000g for 15 min) and incubated in a collagenase/dispase mixture for 1 h at 37°C. Capillary fragments were retained on a 10-μm nylon filter, removed from the filter with endothelial cell basal medium (Life Technologies), supplemented with 20% bovine plasma–derived serum and antibiotics (penicillin 100 units/mL, streptomycin 100 µg/mL), and seeded on 60-mm Petri dishes coated with collagen type IV (5 µg/cm²). Puromycin 3 μg/mL was added for 3 days. Puromycin was then removed from the culture medium and replaced by fibroblast growth factor (2 ng/mL) and hydrocortisone (500 ng/mL).

For transfection, astrocytes were electroporated (2 × 10⁶ astrocytes with 2 µg of plasmid DNA) before seeding using an astrocyte Nucleofector Kit (Amaxa; Lonza, Basel, Switzerland). After electroporation, cells were plated to obtain a final cell density on the day of the experiment similar to that obtained with the transfection method. The transfection efficiency was 60–80%, as assessed with a GFP vector expression.

We used 6-(N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)amino)-6-deoxyglucose (6-NBDG) to measure glucose uptake in astrocytes, as has been shown by others (14). Briefly, cells were starved in serum-free media for 3 h. Then IGF-I (ProSpec-Tany Technogene, Ness Giona, Israel), insulin (Sigma-Aldrich), or vehicle was added to a final concentration of 1 nmol/L. We then added 6-NBDG (Setareh Biotech, Eugene, OR) to a final concentration of 30 μmol/L. Cultures were kept for 3 h at 37°C; then ice-cold PBS was added; cells were trypsinized and collected; and FBS and PBS were added. Fluorescence intensity was measured by flow cytometry (FACSAria Cytometer; BD Biosciences). When pharmacological inhibitors were used, picropodophyllin (120 nmol/L), an IGF-IR–specific inhibitor (picropodophyllotoxin [PPP]; Merck Sharp & Dohme, Hoddesdon, U.K.), and S961 (10 nmol/L), a recently described (15) insulin receptor (IR)–specific inhibitor (Novo Nordisk, Bagsværd, Denmark), were added 1 h prior to the addition of insulin and/or IGF-I.

Glucose levels in astrocytes were determined by the glucose oxidase test (Thermo Scientific) following manufacturer instructions. In brief, astrocyte cultures were washed three times with PBS and starved over 3 h in low-glucose medium without serum; cells were treated with or without IGF-I and insulin over 3 h. After treatments, cells were kept on ice, washed three times with Tris-buffered saline, and lysed with 100 µL of lysis buffer (150 mmol/L NaCl, 20 mmol/L Tris HCl, pH 7.4, 5 mmol/L EDTA, 10% glycerol, and 1% Nonidet P-40). Samples were kept in the freezer at least overnight and centrifuged at 5 min at 4°C and 10,000 rotations/min. Supernatants were collected, and 50 µL was used, as indicated.

Translocation of GLUT1-Flag to the cell membrane was determined by flow cytometry following previously published procedures (16). In brief, cultured astrocytes were labeled with anti-Flag M2 (1:1,000; F1804; Sigma-Aldrich) and secondary antibody Alexa Fluor 488 (1:1,000; A-11008; Life Technologies) and were fixed before assessing fluorescence intensity. In a second type of assay, cell surface proteins were biotinylated following manufacturer instructions (EZ-LinkSulfo-NHS-SS-Biotin; Thermo Scientific). Briefly, astrocytes were washed three times with ice-cold PBS (pH 8.0) and suspended at a concentration of 25 × 10⁶ cells/mL in PBS. Immediately before use, a 10 mmol/L solution of Sulfo-NHS-SS-Biotin was prepared by adding 6 mg/mL ultrapure water, and ∼80 µL of 10 mmol/L Sulfo-NHS-SS-Biotin per mL reaction volume was added and incubated at room temperature for 30 min. Cells were washed three times with ice-cold PBS. Proteins were purified by affinity chromatography using NeutroAvidin Agarose Resin (Thermo Scientific). Biotinylated proteins were resolved by Western blot. The membrane protein Na⁺/K⁺ ATPase was used as a loading control.

GLUT1-IGF-IR interactions were detected in astrocytes grown on glass coverslips using the Duolink II in situ proximity ligation assay (PLA) detection kit (OLink AB, Uppsala, Sweden), as described previously (17). Astrocytes were fixed in 4% paraformaldehyde for 10 min, washed with PBS containing 20 mmol/L glycine to quench the aldehyde groups, permeabilized with the same buffer containing 0.05% Triton X-100 for 5 min, and successively washed with PBS. After immersion for 1 h at 37°C with the blocking solution in a preheated humidity chamber, astrocytes were incubated overnight in antibody diluent medium with the primary antibodies rabbit polyclonal anti-GLUT1 (1:100; catalog #sc-7903; Santa Cruz Biotechnology) and monoclonal mouse anti–IGF-IR (1:100; catalog #sc-463; Santa Cruz Biotechnology), which were processed following the instructions of the supplier using the PLA probes detecting rabbit or mouse antibodies (Duolink II PLA probe anti-Rabbit plus, and Duolink II PLA probe anti-Mouse minus diluted in antibody diluent to a concentration of 1:5) and a DAPI-containing mounting medium. Samples were observed using an SP2 Confocal Microscope (Leica Microsystems, Wetzlar, Germany) equipped with an apochromatic ×63 oil-immersion objective. For images of each field, a maximum projection (superimposed sections) in two channels (one per staining) of 6–12 z stacks with a step size of 1 µm were acquired.

The production of glycogen and lactate in astrocytes was assessed using commercial kits (MAK016 Glycogen Assay Kit; Sigma-Aldrich; MAK064 Lactate Assay Kit; Sigma-Aldrich) following the manufacturer instructions. Briefly, cells treated as described above were kept for 6 h at 37°C. Then, cells were washed with ice-cooled PBS and homogenized in ice water. Samples were measured on a fluorescence microplate reader (Fluostar Optima; BMG LABTECH, Ortenberg, Germany).

Quantitative real-time PCR analysis was carried out as described previously (18). Results were expressed as relative expression ratios on the basis of group means for target transcripts versus reference 18S transcript. At least three independent experiments were performed.

Immunocytochemical assays and Western blotting were performed as described previously (19). Antibodies and dilutions used are shown in Table 1. Densitometric analysis was performed using Analysis Image Program (Bio-Rad, Hercules, CA). A representative blot is shown from a total of at least three independent experiments.

Immunocytochemistry was performed as explained above. Pictures of stained cultures were taken using a Leica TCS-SP5 Confocal Microscope equipped with a ×63 oil-immersion objective (HCX PL APO CS, ×63/1.40–0.60). For each field, a series of 8–12 z stacks with a step size of 0.29 µm in each channel (one per staining) was acquired. The quantification of vesicles in each field was made enhancing the contrast and finally applying a threshold to obtain the binary image and the region of interest around each particle. The number of colocalization particles among the different channels was also determined. The number of particles was determined taking 20–34 cells in five different fields using the Fiji software package (20).

Glucose uptake by astrocytes was evaluated as previously described (21) following procedures that have already been reported (22). The administration of 6-NBDG, the experimental stimulation paradigm, confocal microscopy, and data analysis are described in greater detail elsewhere (23).

¹⁸F-fluorodeoxyglucose (FDG) PET was used to measure brain glucose handling as described (23). Once the ¹⁸F-FDG uptake in the different brain regions was calculated (in kilobecquerels per cubic centimeter), the activity of each left hemisphere region was normalized to its homologous region in the right hemisphere and expressed as proportional uptake (left/right).

Cortical slices were obtained from C57BL/6 mice (12–16 weeks old). Animals were anesthetized and decapitated. The brain was rapidly removed and placed in ice-cold artificial cerebrospinal fluid (ACSF). Slices (400 µm) were incubated for >1 h at room temperature (22–24°C) in ACSF containing the following (in mmol/L): NaCl 124, KCl 2.69, KH₂PO₄ 1.25, MgSO₄ 2, NaHCO₃ 26, CaCl₂ 2, and glucose 10, and gassed with 95% O₂/5% CO₂ (osmolarity 290–300 mOsm/L, pH 7.35). Slices were then transferred to an immersion recording chamber and perfused with gassed ACSF at a rate of 1–2 mL/min and at 30–34°C. The recording area was visualized using an Olympus (Tokyo, Japan) BX50WI Microscope. For electrophysiology, bipolar platinum/iridium electrodes of 50 mm placed on layer 4 of somatosensory cortex. To record excitatory synaptic activity glass microelectrodes of 3–5 mol/LΩ filled with NaCl 1 mmol/L were used and placed in layer 2, and slices were perfused with ACSF containing 50 μmol/L picrotoxin and 5 μmol/L CGP55845 to block GABA_A and GABA_B receptors, respectively, unless otherwise noted. In the ACSF external solution with no added glucose, the concentration of NaCl was adjusted so that the osmolarity becomes 290–300 mOsm/L. A stimulus intensity that evoked half-maximum amplitude field excitatory postsynaptic potentials (fEPSPs) was used. Baseline responses were recorded for at least 10 min with test stimuli given at a rate of 0.1 Hz. The slopes of the baseline responses were set to 100%, and the slopes during the experiments are expressed as percentages of the baseline slope. IGF-IRs and IRs were selectively blocked by PPP (10 μmol/L) and S961 (10 nmol/L), respectively, added to the ACSF. Data are presented as mean ± SEM. The Wilcoxon matched pair test was used to detect any significant changes in synaptic activity by comparing the slopes of the baseline responses with those after the perfusion medium was changed.

Normal distribution tests were performed in all initial sets of experiments, and a nonparametric Wilcoxon test was applied accordingly. For samples with normal distribution, parametric tests include one-way ANOVA followed by a Tukey honestly significant difference test or t test. A P value <0.05 was considered to be significant.

As described over a century ago (24), neuronal activity increases local blood flow through neurovascular coupling at the BBB, and blood nutrients are taken up by the active brain region. Because endothelial cells and astrocytes are the key cell constituents of the BBB involved in blood glucose uptake, we analyzed the effect of I+I on glucose capture by these two types of cells using the fluorescent glucose analog 6-NBDG (25). Whereas neither insulin nor IGF-I modulated glucose uptake in these cells, the combined addition of I+I stimulated it in astrocytes but not in endothelial cells (Fig. 1A and Supplementary Fig. 1A). Glucose uptake by neurons, the major energy consumers of the brain, was not affected by insulin peptides either (Supplementary Fig. 1B). Increased glucose uptake by astrocytes after I+I was reflected by increased glucose content in these cells (Fig. 1B).

Because glucose uptake by astrocytes is stimulated when both I+I are added simultaneously, an interaction of both hormones through their respective receptors was deemed probable. We confirmed by quantitative PCR (data not shown) that insulin and IGF-IRs are expressed by astrocytes (26,27) and inhibited them with drug antagonists. We found that either S961, an inhibitor of IR, or PPP, an inhibitor of IGF-IR, blocked the effects of I+I (Fig. 1C and D). In addition, the expression of dominant-negative forms of either IGF-IR or IR abolished the stimulatory actions of I+I (Fig. 1E and F). Reduced IR activity had a pronounced effect on astrocytic basal glucose uptake (Fig. 1E). Therefore, both receptors are required to stimulate astrocyte glucose uptake because inhibiting either one is sufficient to block the effect of I+I.

Because blood-borne glucose is captured by the brain through GLUT1, the most abundant type of facilitative transporter in brain endothelia and astrocytes (28), we determined whether this transporter is involved in glucose uptake after I+I. Indeed, the amount of GLUT1 translocated to the astrocyte cell membrane was significantly increased in the presence of I+I, as determined by flow cytometry of tagged GLUT1 (Fig. 2A and Supplementary Fig. 1C), or by biotinylation of membrane proteins (Fig. 2B), whereas GLUT1 immunostaining in astrocytes was redistributed to the cell membrane only in the presence of I+I (Supplementary Fig. 1D), suggesting its mobilization from internal stores. Furthermore, when GLUT1 levels in astrocytes were reduced with shRNA (Supplementary Fig. 2A), basal glucose uptake was drastically reduced and the stimulatory action of I+I no longer seen (Fig. 2C). Because glucose uptake in astrocytes is associated with gluconeogenesis (29), we determined glycogen levels in astrocytes exposed to I+I and found them significantly elevated (Fig. 2D), whereas the production of lactate, a major glucose metabolite in astrocytes, was unaffected (Supplementary Fig. 2B). As previously reported, insulin also increased glycogen levels without affecting lactate production (30).

Although it is widely assumed that pancreatic insulin underlies the brain actions of this hormone (31), it is now firmly established that the brain also produces insulin (32). The same is true for IGF-I, a hormone that reaches the brain from the circulation (33) but is also locally produced (34). To determine whether locally produced I+I participate in brain glucose metabolism, we analyzed the effect of the blockade of IRs/IGF-IRs in neuronal activity using brain slices under changing energy demands. Importantly, both insulin and IGF-I mRNA are detected in brain slices used for these experiments (Supplementary Fig. 2C). We took advantage of the fact that during high energy needs neurons rely on astrocytes for energy supply (35) and that neuronal activity triggers both insulin release (36) and IGF-I release (37). We submitted cortical slices to hypoglycemia in the presence of drugs blocking IRs/IGF-IRs. As expected, fEPSPs decayed after glucose withdrawal (Fig. 3A) and glucose readdition to the medium triggered the recovery of fEPSPs. However, in the presence of either PPP or S961, recovery was significantly reduced (Fig. 3A).

We next determined whether astrocytes are involved in the actions of I+I on neuronal activity. For these experiments, we took advantage of two previous observations: 1) upon sensory stimulation, glucose uptake in somatosensory cortex is predominantly mediated by astrocytes (21); and 2) direct application of the IGF-IR blocker PPP to the somatosensory cortex attenuates neuronal firing elicited by sensory stimulation (19). In this regard, it is important to note that stimulation of the whiskers activates both IGF-IR and IRs in somatosensory cortex (Supplementary Fig. 3A). Using in vivo fluorescence microscopy of the mouse somatosensory cortex and whisker stimulation (21), we observed reduced glucose uptake (detected with the green fluorescent marker 6-NBDG) in astrocytes (identified with the astrocyte-specific fluorescent red marker SR101) of mice receiving a local infusion of PPP compared with saline-infused mice (Fig. 3).

Intracellular signaling pathways downstream of cooperative actions of I+I were then examined. We measured phosphorylation of AKT and mitogen-activated protein kinase (MAPK) because these two kinases form part of the canonical signaling of both IRs and IGF-IRs. As shown in Fig. 4A, phospho-MAPK, but not phospho-AKT (Supplementary Fig. 3B) was synergistically elevated by I+I in astrocytes. As potential downstream mechanisms, we focused on protein kinase D (PKD [also known as PKCμ]), a diacylglycerol-binding kinase that may act downstream of MAPKs (38). PKD regulates the translocation of GLUTs (39) and is specifically associated with IGF-IR, but not with IR (40) (see below). PKD was synergistically stimulated by I+I, as documented by increased levels of phospho-Ser-PKD (Fig. 4B), and in a MAPK-dependent manner (Fig. 4C). Further, the PKD inhibitor CID755673 (CID) abrogated increased glucose uptake by astrocytes in response to I+I (Fig. 4D). To determine the in vivo significance of these findings, we analyzed brain glucose uptake in mice lacking PKD in GFAP astrocytes (Supplementary Fig. 4A–D). Glucose uptake in response to whisker stimulation in somatosensory cortex was significantly more increased in PKD1^floxed littermates than in PKD1^Δ animals (Fig. 4E). Collectively, these data support a cooperative action of I+I in glucose handling by astrocytes through the synergistic activation of MAPK_42–44 and PKD.

Since GLUT1 is involved in glucose uptake promoted by I+I, we examined the mechanisms linking MAPK/PKD activation with its translocation to the cell membrane. GIPC (41,42) is a scaffolding protein that participates in the trafficking of GLUT1 (42). We observed that GIPC interacts with GLUT1 in response to I+I (Fig. 5A). Furthermore, GIPC also interacts with IGF-IR, but not with IR, after I+I (Fig. 5A). Because GIPC modulates the activity of the small GTPase RAC1 (43), which is involved in the translocation of GLUTs from intracellular compartments to the cell membrane (44), we also analyzed its possible involvement in the action of I+I. We found that I+I promoted the interaction of RAC1 with GIPC, GLUT, IR, and IGF-IR (Fig. 5A). Reversed coimmunoprecipitation using anti-RAC1 or anti–IGF-IR antibodies confirmed these interactions (Fig. 5B and C). Further, the activity of RAC1 was increased after I+I administration, although IGF-I also increased it when given alone (Fig. 5D). To link the observed protein-protein interactions promoted by I+I with activation of PKD by I+I, we analyzed whether the inhibition of PKD with CID altered these interactions. Indeed, CID markedly decreased the association of RAC1 with GLUT1 after I+I administration (Fig. 5E), or the interaction of GLUT1 with IGF-IR, as measured by PLAs (Fig. 5F).

Using the immunocytochemistry of GIPC, IGF-IR, and GLUT1 in astrocytes, we confirmed that I+I increases interactions among these proteins. The addition of insulin greatly decreased all types of immunostained puncta, and IGF-I slightly modified some of them; whereas I+I synergistically increased the interactions of IGF-IR/ GLUT1, IGF-IR/GIPC, and GLUT1/GIPC (Fig. 6A and B; P < 0.05 vs. basal conditions). To link the pattern of puncta mobilization after I+I and glucose uptake, we inhibited the latter by blocking PKD with CID (Fig. 4D). CID drastically altered the pattern of puncta mobilization elicited by I+I (Fig. 6B), further supporting a role for this pattern of protein-protein interactions in glucose uptake. To firmly establish a role of GIPC in the interaction between IGF-IR and GLUT1, we reduced GIPC levels in astrocytes using GIPC shRNA (Supplementary Fig. 4E). GIPC is required to sort GLUT1 to the plasma membrane (45). Accordingly, GIPC reduction decreased the amount of GLUT1 in the cell membrane (Fig. 6C), and as a result, basal glucose uptake was reduced and stimulation by I+I impaired (Fig. 6D).

These results present evidence of a cooperative mechanism between insulin and IGF-I for glucose uptake by astrocytes of consequence on brain glucose handling. The process invokes an interaction of I+I in active brain regions to synergistically stimulate MAPK/PKD, which in turn promotes the translocation of GLUT1 to the cell membrane. The latter involves the interaction of GLUT1 with the scaffolding protein GIPC, which also interacts with the IGF-IR, and the GTPase RAC1 that is activated and interacts with both the IGF-IR and IR (Supplementary Fig. 4F). The presence of additional protein-protein interactions is likely and will require further study.

The role of insulin in the regulation of glucose handling by the brain remains elusive (46), because under physiological circumstances it does not seem to modulate it (1); whereas, in pathologies such as diabetes, brain glucose metabolism is modulated by insulin (2), and, in normal brains, hippocampal GLUT4 is translocated to the cell membrane in response to this hormone (47). It is possible that our understanding of the role of this hormone in brain glucose metabolism has been hindered by the fact that a concerted action with IGF-I seems to be required. Because astrocytes may be involved in energy allocation to neurons through the so-called “lactate shuttle” (48), our experiments in brain slices suggest that local I+I regulate neuronal function under changing energy demands, probably by modulating astrocyte support. Favoring the latter is the observation that the blockade of IGF-IRs in somatosensory cortex inhibits both neuronal activity in response to sensory stimulation (19) and glucose uptake by astrocytes (Fig. 3B and C). The experiments using brain slices also favor the notion that local I+I is sufficient to modulate glucose handling. The fact that brain insulin may modulate glucose handling by this organ, in concert with brain IGF-I, may explain the difficulty in determining a role for circulating insulin in brain glucose handling. Because diabetes impacts the brain at biochemical, structural, and behavioral levels (6), it is possible that the action of brain insulin may be also altered, unveiling a role for circulating insulin in the diabetic state (2). Furthermore, a role for circulating I+I cannot be entirely ruled out yet, because, at least for the latter, neuronal activity triggers its entrance into the brain (19). Thus, while we still do not know whether circulating insulin enters the brain in response to neuronal activation, it is possible that both insulin and IGF-I increase locally on demand and may in this way contribute to glucose uptake by astrocytes located in the vicinity of active neurons.

The fact that insulin acts in concert with IGF-I to modulate GLUT1 may also imply a role of the latter in the interaction with other energy substrates, such as ketone bodies and amino acids. Hence, not only insulin deficiency but also IGF-I deficits may potentially impact these pathways in the face of increasing energy demands, such as during hypoglycemia. These intriguing possibilities also require further analysis.

Collectively, these results indicate that an interactive network of I+I plays an essential role in glucose handling by astrocytes impacting in brain glucose metabolism. These observations may open a new path for future therapeutic approaches for brain complications in diabetes (49) and for diabetes-like pathology in neurodegenerative diseases such as Alzheimer dementia (50).

Acknowledgments. The authors thank M. Garcia (Cajal Institute) and L. Guinea (Cajal Institute) for technical support. The authors also thank Novo Nordisk (Denmark) for providing S961 insulin receptor inhibitor.

Funding. This work was funded by MINECO (Spain) grants SAF2010-60051 and SAF2013-40710-R, and by CIBERNED. E.H.-G. was partially funded by a fellowship from ColFuturo and CIBERNED. T.I. is funded by MINECO (SAF2014-52737-P). T.I. and I.T.A. are funded by CIBERNED and Comunidad de Madrid, Spain (P2010/BMD-2331-Neurodegmodels).

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Authors Contributions. A.M.F. designed and performed in vivo experiments and analyzed the results. E.H.-G. designed and performed in vivo and in vitro experiments and analyzed the results. P.P.-P. designed and performed in vitro experiments and analyzed the results. A.P.-A. performed confocal microscopy experiments and analyzed the results. S.M. performed slice experiments. T.M. performed brain glycogen experiments and analyzed the results. A.S., A.T.-S., and L.G.-G. performed in vitro experiments in astrocytes. J.P.-U. characterized in vitro PKD1 astrocytes. J.F. and E.N.O. designed and obtained PKD floxed mice. R.F.d.l.R. performed PET experiments. L.G.G. characterized PKD1 mice brain tissue and performed PET experiments. M.A.P. analyzed PET results. T.I. designed and obtained PKD-KO mice. A.A. designed confocal microscopy experiments. H.S. designed brain glycogen experiments and analyzed the results. G.P. performed and analyzed the slice experiments. E.D.M. designed and performed confocal experiments. I.T.A. designed the study and wrote the manuscript. I.T.A. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. A preliminary version of this work was published in BioRxiv (available at http://biorxiv.org/content/early/2015/07/31/023556).