GLP-1 and Exendin-4 Transiently Enhance GABA_A Receptor–Mediated Synaptic and Tonic Currents in Rat Hippocampal CA3 Pyramidal Neurons

In recent years, compelling evidence has emerged suggesting that diabetes mellitus increases the risk for cognitive impairments in the elderly (1–8). How this comes about is not resolved, but interestingly, the brain contains receptors for many metabolic hormones, among those receptors for insulin and the incretins. To date, with the exception of the hypothalamus, we know relatively little about how metabolic hormones affect neuronal activity and thereby brain function. The hippocampus is central for cognitive functions and is the center for memory and learning (9,10). It has prominent expression for receptors activated by metabolic hormones (10). Furthermore, via neurons in the septum, the hippocampus regulates the activity of a number of hypothalamic nuclei (11,12). Glucagon-like peptide-1 (GLP-1) is a gut hormone that is secreted by intestinal L cells in response to food intake, and the GLP-1 receptor is expressed in the hippocampus (10,13). GLP-1 crosses the blood-brain barrier, but it is also a neurotransmitter produced by neurons with cell bodies in the brainstem (12–15). The best known effects of GLP-1 are to stimulate insulin and inhibit glucagon secretion in a glucose-dependent manner in the pancreatic islets to regulate glucose homeostasis after a meal (13). Although the GLP-1 receptor is expressed in the hippocampus (10,16,17) and GLP-1 and its mimetics, e.g., exendin-4, liraglutide, might potentially be used to treat cognitive declines related to diabetes (6,7), to date not much is known about the effects of GLP-1 on neuronal signaling and, hence, how GLP-1 affects cognition and hippocampal regulation of metabolic homeostasis.

The GLP-1 is a 30–amino acid–long peptide and is derived from posttranslational processing of the preproglucagon gene (18). Initially, the peptide GLP-1 (1-37) was identified from this processing, but later it was shown that there were two shorter peptides, GLP-1 (7-37) and GLP-1 (7-36)amide, that were the active species in vivo. The half-life of the peptides in plasma is very short, only about 1–2 min (13,19), due to degradation by the enzyme dipeptidyl peptidase-4 (13). In the pancreatic islets, the GLP-1 receptor is internalized after GLP-1–induced activation (20–23) and passes through recycling endosomes before it appears in the plasma membrane again (23). The GLP-1 receptor is widely distributed in the brain (10,16,24), including in the hippocampal CA3 pyramidal neurons (16,25), and GLP-1 and other agonists at the GLP-1 receptor have been reported to regulate food intake (26), be neuroprotective (27), anti-inflammatory (28), and modulate synaptic plasticity and memory formation (28–32).

γ-Aminobutyric acid (GABA), the major inhibitory neurotransmitter in the central nervous system, activates synaptic and extrasynaptic GABA_A receptors that mediate synaptic and tonic currents, respectively, regulating activity of neurons and neuronal circuits (33). Metabolic hormones are emerging as modulators of GABA signaling in hippocampal neurons. Already in 1984, Palovcik et al. (34) demonstrated that insulin inhibits pyramidal neurons in hippocampal slices. Later, insulin was shown to enhance miniature inhibitory postsynaptic currents (mIPSCs) in cultured hippocampal neurons (35), and recently we demonstrated that insulin turns on high-affinity GABA_A receptors that generate tonic currents in hippocampal CA1 pyramidal neurons in rat brain slices (36). In the current study, we examined the effects of GLP-1 and exendin-4 on GABA_A signaling in hippocampal CA3 pyramidal neurons in rat brain slices. We found that low physiological GLP-1 concentrations (pico- to nanomoles per liter) and clinically relevant exendin-4 concentrations transiently potentiate synaptic and tonic GABA-activated currents.

Brain slices were dissected for electrophysiological recordings from 16- to 22-day-old Wistar rats. All animal procedures were conducted in accordance with the local ethics guidelines and approved animal care protocols by the Uppsala djurförsöksetiska nämnd, Uppsala, Sweden (Uppsala Animal Ethical Board). Hippocampal slices were prepared as previously described (37). Briefly, the animal was decapitated and the brain rapidly removed and immersed into the ice-cold artificial cerebrospinal fluid (ACSF) containing (in millimoles per liter): 124 NaCl, 3 KCl, 2.5 CaCl₂, 1.3 MgSO₄, 26 NaHCO₃, 2.5 Na₂HPO₄, and 10 glucose with pH 7.3–7.4 when bubbled with 95% O₂ and 5% CO₂. Sagittal or coronal hippocampal slices 400 µm thick were prepared with a vibratome (Leica VT1200S) in the ice-cold ACSF gassed with 95% O₂ and 5% CO₂. Slices were incubated in the same ACSF at 37°C for 1 h and kept at room temperature (20–22°C) during experiments.

All patch-clamp recordings were performed at room temperature (20–22°C). Drugs were in general purchased from Sigma-Aldrich (Germany) or Anaspec [GLP-1, exendin-4, and exendin (9-39)]. Bicuculline methiodide from Santa Cruz Biotechnology (Heidelberg, Germany) or Sigma-Aldrich (Schnelldorf, Germany) was used. The pipette solution contained (in millimoles per liter): 140 CsCl, 1 CaCl₂, 3 EGTA, 0.5 KCl, 1 MgCl₂, 2 ATP-Mg, 0.3 GTP-Na, 5 QX-314 bromide, and 10 TES, pH of 7.25 with CsOH. In some experiments, an inhibitor of the GABA_B receptor, CGP52432 (5 μmol/L), was used, but it did not change the results. The recording pipettes were made from borosilicate glass capillaries (Harvard Apparatus UK) with DMZ-Universal Puller (Zeitz Instruments; Martinsried, Germany) and had resistance of 2–4 MΩ when filled with the pipette solution. The holding potential (V_h) was set to –60 mV and used in all experiments. ACSF, containing kynurenic acid (3 mmol/L) and other drugs, was continuously perfused (3 mL/min) through the recording chamber during experiments. Patch-clamp recordings were done using an Axopatch 200B amplifier (Molecular Devices), filtered at 2 kHz, sampled at 10 kHz by analog-to-digital converter Digidata 1322A (Molecular Devices), and stored in a computer. The recordings were analyzed with pClamp 10 (Molecular Devices) and MiniAnalysis 6 (Synaptosoft, Inc.) software. The amplitude of the tonic current was defined as the difference between the baseline current levels before and after the drug application (38) and the frequency of the spontaneous IPSCs (sIPSCs) immediately before first drug application was defined as control. The maximal drug effect on the sIPSC frequency was calculated and normalized to its control value in the same cell. The average value of the baseline current during the transient change in the current value during GLP-1 application was fitted with a double exponential function: y = y₀ + A₁ × exp(−t/τ_rise) − A₂ × exp(−t/τ_decay), where y₀ and A_1,2 are arbitrary constants and the τ_rise/decay are time constants for the rise and the decay phase of the transient current, respectively.

Total RNA was isolated from rat hippocampal slices by using a GenElute Mammalian Total RNA Miniprep kit (Sigma-Aldrich) and quantified with Nanodrop (Nanodrop Technologies). Rat hippocampal total RNA (100 ng) was reverse transcribed into cDNA in a 20-μL reaction mixture using Superscript III reverse transcriptase (Invitrogen). Negative control was performed by omitting reverse transcriptase in the reaction in order to confirm no genomic DNA contamination in the isolated RNA. Human hippocampal cDNA was purchased from USBiological. PCRs were done in a 10-μL reaction mixture containing 4 μL cDNA (4 ng), 1× PCR buffer, 3 mmol/L MgCl₂, 0.3 mmol/L deoxyribonucleotide triphosphate, 1× carboxy-X-rhodamine reference dye, 0.7 units JumpStart Taq DNA polymerase (Sigma-Aldrich), 0.5× SYBR Green I (Invitrogen), and 0.4 μmol/L each of forward and reverse primers. The primer pairs were synthesized by Sigma-Aldrich: rat Glp1r (forward, GGCATTGTCAAGTATCTCTAC; reverse, GATGAAGACAAGGAAGTTGAC; amplicon size, 123 bp) and human GLP1R (forward, ACATCAAATGCAGACTTGCC; reverse, TCACAAAGGCAAAGATGACC; amplicon size, 81 bp). Amplification was performed in 384-well optical plates using the ABI PRISM 7900HT Sequence Detection System (Applied Biosystems) with an initial denaturation of 5 min at 95°C, followed by 45 cycles of 95°C for 15 s, 60°C for 30 s, and 72°C for 30 s. A melting curve was determined at the end of cycling to ensure the amplification of a single PCR product. Five microliters of each individual PCR were then electrophorized on a 2% agarose gel stained with SYBR Gold (Invitrogen, Carlsbad, CA).

Statistical analysis was carried out using SigmaPlot 11 (Systat Software), MiniAnalysis 6 (Synaptosoft), or GraphPad Prism 6 software. Results are presented as means ± SEM. Paired Student t test was used for data sets normally distributed. The Tukey method was used to detect the outliers. Statistical analysis was performed after excluding the outliers. Nonparametric Mann-Whitney test was used for data sets that were not normally distributed. One-way ANOVA Bonferroni post hoc test was used for multiple comparisons with normally distributed data.

The GABA_A-mediated whole-cell currents were recorded from rat hippocampal CA3 pyramidal neurons bathed in ACSF in the presence of kynurenic acid. At the end of every experiment, bicuculline (100 µmol/L), a GABA_A receptor antagonist, was applied to block GABA-evoked currents. The sIPSCs were abolished by 100 µmol/L bicuculline (Fig. 1) and the holding current decreased, revealing the prominent tonic GABA-activated current normally present in the hippocampal CA3 pyramidal neurons (24.7 ± 1.5 pA, n = 19) (Fig. 1). We then proceeded and examined the effects of GLP-1 on these GABA_A receptor–mediated currents.

We examined whether GLP-1 concentrations ranging from 10 pmol/L to 10 nmol/L affected the GABA-evoked currents, and representative results are shown for three cells in Fig. 2A–C. The GLP-1 receptor mRNA is expressed in both human and rat hippocampus (Fig. 2D). GLP-1 in a concentration-dependent manner transiently enhanced the synaptic and tonic GABA-activated currents in the neurons. The 10 pmol/L GLP-1 concentration was most effective and transiently increased the most frequent sIPSC amplitude by a factor of 1.8 (Fig. 3A), and the average sIPSC frequency was similarly increased (Fig. 3B) by a factor of 1.8 (n = 7) compared with control. Higher GLP-1 concentrations (100 pmol/L, 1 nmol/L, and 10 nmol/L) (Fig. 3A) did not increase the most frequent sIPSC amplitude. However, in 100 pmol/L and 1 nmol/L GLP-1, the average sIPSC frequency still increased by a factor of 1.6 (n = 6) and 1.8 (n = 5), respectively, compared with control (Fig. 3B). The tonic current in the CA3 pyramidal neurons increased when exposed to 10 pmol/L to 1 nmol/L concentrations of GLP-1 but was similar to control in 10 nmol/L GLP-1 (Fig. 3C) resulting in a bell-shaped-like concentration-response relationship. When several GLP-1 concentrations were sequentially applied to a neuron, then each new application led to a transient increase of the tonic current, which subsequently relaxed to the initial current level. Thus, the tonic current amplitude changed transiently in a GLP-1 concentration–dependent manner. Just applying the extracellular solution did not induce the transient increase in the tonic current amplitude. The amplitude was not related to whether the response was obtained after a single exposure to GLP-1 (Fig. 3C) or whether the neuron had previously been exposed to another concentration of GLP-1 (Fig. 3C).

The time course of the tonic current during the first minutes of the GLP-1 application was U shaped (Fig. 2A–C and 4A and B) and could be fitted with a double exponential function: y = y₀ + A₁ × exp(−t/τ_rise) − A₂ × exp(−t/τ_decay) where y₀ is the initial baseline current value and A_1,2 are arbitrary constants but τ_rise and τ_decay are the time constants for the rise and the decay phase of the transient current, respectively. At all GLP-1 concentrations, the current increased with a characteristic time constant of ~2 min (τ_rise), and after an additional 2 min (τ_decay) (Fig. 4C) the current had returned to approximately the initial current level. Both the sIPSCs and the tonic currents in the presence of GLP-1 were blocked by 100 μmol/L bicuculline (Fig. 2A–C).

We examined whether the effects of GLP-1 on the GABA-activated synaptic and tonic currents could be prevented by a GLP-1 receptor antagonist. Exendin (9-39) (Ex9-39) is a competitive inhibitor of GLP-1 at the GLP-1 receptor (23). Since GLP-1 once bound to the GLP-1 receptor starts intracellular cascades leading to activation of various proteins, it is essential to apply the inhibitor first and only then coapply GLP-1 together with the inhibitor in order to prevent GLP-1 effects on neuronal function such as modulation of the GABA signaling. In our experiments, we therefore applied Ex9-39 first and then coapplied GLP-1 together with Ex9-39 to the brain slices. When Ex9-39 was applied to the hippocampal slices, it inhibited the effects of 10 pmol/L GLP-1 on the GABA-activated currents but had no effects when applied alone (Fig. 5). The amplitude and frequency of the synaptic currents were now similar to control currents (Fig. 5A–C), and there was no increase in the amplitude of the tonic current (Fig. 5A and D).

In order to examine whether the GLP-1 effects on the currents were due to pre- or postsynaptic mechanisms or both, we studied the influence of GLP-1 on the currents in the presence of the voltage-gated sodium channel blocker tetrodotoxin (TTX) (1 µmol/L). TTX inhibits action potential–dependent GABA release. The amplitudes of the synaptic currents were similar in TTX and TTX plus 10 pmol/L GLP-1 (Fig. 6A and B). The results further show that the frequency of the IPSCs decreased when the slices were exposed to TTX from 16.8 ± 1.2 to 2.3 ± 0.4 Hz (nonparametric Mann-Whitney test, P = 0.008, n = 5), respectively, and remained at a similar level in the presence of 10 pmol/L GLP-1 plus TTX (Fig. 6A and C). The results are consistent with GLP-1 potentiating the release of GABA from presynaptic terminals. The tonic current also decreased in TTX compared with control, but in contrast to the synaptic currents, the effect of GLP-1 on the tonic current was maintained (Fig. 6A and D). The tonic current transiently increased from 11.1 ± 2 pA to 22.7 ± 2.3 pA (n = 6) when the slices were exposed to GLP-1 (10 pmol/L) in the presence of TTX (Fig. 6A and D). These results demonstrate that GLP-1 signaling modulates high-affinity, extrasynaptic GABA_A receptors in the plasma membrane of CA3 pyramidal neurons that generate the tonic current.

Exendin-4 is an agonist at the GLP-1 receptors. We therefore examined whether exendin-4 had effects similar to those of GLP-1 on the synaptic and tonic GABA_A receptor–mediated currents in the hippocampal CA3 pyramidal neurons. Representative results for 10, 50, and 100 nmol/L exendin-4 are shown for one cell in Fig. 7A. In a concentration-dependent manner, exendin-4 transiently enhanced both the synaptic and tonic GABA-activated currents in the neurons. Exendin-4 did not increase the most frequent sIPSC amplitude (Fig. 7B), whereas the average sIPSC frequency (Fig. 7C) was enhanced by a factor of 1.4 (n = 6), 1.5 (n = 6), and 1.4 (n = 6) by 10, 50, and 100 nmol/L exendin-4, respectively, but did not change in 0.5 nmol/L exendin-4. All exendin-4 concentrations tested (0.5, 10, 50, and 100 nmol/L) transiently enhanced the tonic current in the CA3 pyramidal neurons (Fig. 8A). The time course of the tonic current enhancement by exendin-4 was U shaped. It was similar to what was recorded in GLP-1 (Fig. 7A and 8B and C) and could be fitted by equation 1 [y = y₀ + A₁ × exp(−t/τ_rise) − A₂ × exp(−t/τ_decay)]. Similar to tonic currents evoked by GLP-1, for all exendin-4 concentrations, the currents increased with a characteristic time constant of ~2 min (τ_rise) (Fig. 8D), and after an additional 2 min (τ_decay) (Fig. 8D), the current had returned to approximately the initial current level. Both the sIPSCs and the tonic currents in the presence of exendin-4 were inhibited by 100 μmol/L bicuculline (Fig. 7A).

Similar to GLP-1, effects of exendin-4 on the GABA_A receptor–activated currents might be related to either pre- or postsynaptic mechanisms or both. We therefore examined the effects of exendin-4 on the currents in the presence of TTX (1 μmol/L) that inhibits action potential–dependent transmitter release (Fig. 9A–D). The slices were first exposed to TTX to block presynaptic GABA release and then exposed to 10 nmol/L exendin-4 (Fig. 9A). The amplitudes of the synaptic currents were similar in TTX alone or TTX plus exendin-4 (Fig. 9B), but the frequency of the IPCSs decreased to 2.2 ± 0.4 Hz when the slices were exposed to TTX and remained at a similar level in the presence of 10 nmol/L exendin-4 plus TTX (n = 4) (Fig. 9C). At the same time, exendin-4 enhanced the tonic current in the presence of TTX (Fig. 9D). The results are consistent with exendin-4 enhancing GABA release from presynaptic terminals and modulating tonic currents in the postsynaptic neurons and are similar to the results obtained with GLP-1.

The metabolic hormone GLP-1 and its mimetic exendin-4 both enhanced GABA signaling in rat hippocampal CA3 pyramidal neurons. The CA3 pyramidal neurons are regulated by a number of inhibitory interneurons and form an important part of the hippocampal neuronal circuit involved in memory formation (9). The synaptic and tonic GABA-activated currents in the CA3 pyramidal neurons were transiently enhanced by GLP-1 at physiologically relevant concentrations and by exendin-4 at concentrations relevant when treating type 2 diabetes. The effects of GLP-1 and exendin-4 on the GABA signaling can be attributed to both presynaptic and postsynaptic mechanisms. The results are consistent with GLP-1 modulating cognitive process in a hippocampal dependent manner.

There appears to be a number of ways GLP-1 and exendin-4 can enhance GABA signaling in the CA3 pyramidal neurons. The most frequent sIPSC amplitude was approximately doubled and the tonic current increased by 60% in 10 pmol/L GLP-1. These increases potentially can be attributed to a number of processes such as increased release of GABA from the presynaptic terminal, increased number of GABA_A receptors in the postsynaptic membrane, increased spillover of GABA from the synaptic cleft, or insertion of new or modified GABA_A receptors with higher affinity in the postsynaptic membrane (33,35,36,39). TTX inhibits voltage-gated sodium channels. In solutions containing TTX, action potential generation is inhibited resulting in decreased transmitter release from presynaptic terminals. We used TTX to differentiate between presynaptic and postsynaptic effects of GLP-1 and exendin-4 on the GABA signaling. In the presence of TTX, GLP-1 affected neither the amplitudes nor the frequency of the synaptic currents, and similarly, exendin-4 in the presence of TTX no longer modulated the frequency of the IPSCs. The enhancing effect of GLP-1 or exendin-4 on the sIPSC frequency in the absence of TTX is, therefore, related to increased release of GABA from presynaptic terminals. The effects on the tonic current were more complex. In the presence of TTX, the tonic current was reduced by ~50%. This is in accordance with at least a part of the tonic current amplitude being related to the magnitude of spillover of GABA from synapses (33,39). However, the remaining tonic current recorded in TTX was still sensitive to GLP-1 and approximately doubled in amplitude when the neurons were exposed to 10 pmol/L GLP-1, and similar results were obtained with exendin-4. These results are consistent with GLP-1 and exendin-4 modulating GABA signaling in the hippocampal CA3 neuron in a postsynaptic manner. The GABA_A receptors generating the tonic current and modulated by GLP-1 receptor signaling in the postsynaptic neuron are high-affinity, extrasynaptic GABA_A receptors that are activated by the very low, ambient GABA concentrations present around the neurons. Where the GABA originates from is not clear, but mechanisms involving nonvesicular release (40–42) such as reversal of GABA transporters or release of GABA from astrocytes have been proposed. That pre- and postsynaptic mechanisms can regulate tonic GABA_A receptor–mediated currents in hippocampal neurons is in accordance with previous reports (33,35,36,39). Interestingly, we previously reported that another metabolic hormone, insulin, can modulate high-affinity GABA_A receptors in hippocampal CA1 pyramidal neurons (36). Axons from CA3 pyramidal neurons, the Schaffer collaterals, project to the CA1 pyramidal neurons where they synapse. Both the CA3 and the CA1 neurons have critical roles in hippocampal-dependent memory and learning processes (43).

Metabolic hormones are emerging as important regulators of hippocampal function (28,44–47). GLP-1 has previously been shown to decrease glutamate-generated currents in cultured hippocampal neurons (48) and decrease hippocampal θ wave duration in rats (30), and there is an impairment of synaptic plasticity and memory formation in GLP-1 receptor knockout mice (49). The hippocampus is well-known for its role in memory encoding, but less focus has been on its function in governing body physiology (10,11). The hippocampus contains receptors for molecules regulating many physiological processes including receptors for the metabolic hormones (10). Furthermore, via neurons in the septum, the hippocampus maps in a topographical manner onto the hypothalamus and generally results in inhibition of hypothalamic activity (10,11). In hippocampal CA1 neurons, insulin enhances IPSCs (35) and insulin also turns on high-affinity GABA_A receptors generating tonic currents in these neurons (36), which normally express very small or no tonic currents (33). GLP-1 is made in the brain stem and by L cells in the intestine and crosses the blood-brain barrier (13). In plasma, the half-life of GLP-1 is 1–2 min, and the maximal concentration after a meal is <40 pmol/L (13). Interestingly, in our experiments, 10 pmol/L GLP-1 and exendin-4 effectively enhanced the GABA-activated current response with a similar time constant of activation (2 min), an apparent synchrony with the lifetime of GLP-1 in plasma. Recently, Roed et al. (23) determined the half-maximal concentration for activation (EC₅₀) of the GLP-1 receptor expressed in human embryonic kidney cells to be 9.8 ± 1.0 pmol/L. They further determined the EC₅₀ of GLP-1 for inducing GLP-1 receptors internalization to be 12 ± 5 nmol/L and showed that maximal internalization level occurred with supersaturating concentrations (1 μmol/L) within 15–20 min. In the CA3 neurons in the presence of an inhibitor of the GLP-1 receptors, Ex9-39, no effects of GLP-1 on the synaptic or tonic GABA-activated currents were recorded. The activation and decay phase time constants of 2 min for the tonic current were concentration independent—similar for the two agonists and much faster than the reported rate for GLP-1 receptors internalization. The time constant of the tonic currents rising phase reflects the activation of GLP-1 receptors by the agonists, GLP-1 or exendin-4, transduction of the signal inside the neuron, and as a result, increased current through the extrasynaptic GABA_A receptors in the postsynaptic plasma membrane. What the decay time constant of the tonic current represents is unclear. However, as the effect of both GLP-1 and exendin-4 on the synaptic currents was transient, it is possible that the decay of the responses is related to desensitization of the GLP-1 receptors (50).

Midlife type 2 diabetes increases the likelihood for cognitive decline and brain atrophy (1,2). A number of studies published to date indicate that diabetes and poor glycemic control are significant risk factors for decreased memory functions later in life (1–8). GLP-1 receptor agonists, e.g., exendin-4, liraglutide, are reported to be neuroprotective and possible rescue cognition in models of Alzheimer, Parkinson, and Huntington disease (7,51). Cognitive decline appears to be a significant complication associated with diabetes, and it is therefore of major interest to characterize the effects of molecules activating the GLP-1 receptor in hippocampal brain tissue, as these compounds may potentially reverse or slow down the decline. Our results show that exendin-4 application effectively mimics the GLP-1 effects on the GABA_A signaling in the CA3 pyramidal neurons, but the modulation varies somewhat between the two GLP-1 receptor agonists, e.g., the sIPSC amplitudes were not enhanced by any of exendin-4 concentrations that we used in this study, whereas the sIPSC amplitudes were enhanced by 10 pmol/L GLP-1.

In conclusion, the results demonstrate that both GLP-1 and exendin-4 effectively potentiate GABAergic signaling in hippocampal CA3 pyramidal neurons. Figure 9E shows a cartoon with pre- and postsynaptic neuronal terminals and the synapse and identifies the location of the different receptors. It also shows where the GABA-evoked currents that are modified by the GLP-1 receptor agonists are generated. At physiological GLP-1 concentrations, both synaptic and tonic GABA-activated currents were transiently enhanced, and the effects were blocked by Ex9-39. The results demonstrate that GLP-1 and exendin-4 modulate GABA signaling in hippocampal neurons in both pre- and postsynaptic manners. The results imply that in order to combat diabetes-associated cognitive dysfunction with GLP-1 or medicines that mimic GLP-1 actions, e.g., exendin-4, a good understanding of GLP-1 receptor agonist effects on hippocampal neuronal functions is desirable. Furthermore, hippocampal-dependent learning and memory mechanisms may potentially contribute directly to control of energy homeostasis by regulating hypothalamic function. In order to elucidate the interplay between cognitive function and diabetes, more studies of regulation of hippocampal function by metabolic hormones are required.

Acknowledgments. The authors thank Amol Bhandage and Frida Lindberg (Neuroscience, Uppsala University) for help with the RT-PCR experiments.

Funding. This work was supported by a grant from the Swedish Research Council to B.B. and a grant from Åke Wibergs Foundation to Z.J. S.V.K. held a postdoctoral fellowship from EXODIAB, and Z.J. held a postdoctoral fellowship from M.Å.H. Ländells foundation and M. Sjöströms Foundation.

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

Author Contributions. S.V.K. designed the experiments, performed the experiments, analyzed data, made the figures, performed the statistical analysis, and wrote the manuscript. Z.J. designed the experiments, performed the experiments, and edited the manuscript. O.B. performed the experiments. B.B. designed the experiments and wrote the manuscript. B.B. 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.