Lateral hypothalamus hypocretin/orexin glucose-inhibited neurons promote food seeking after calorie restriction

Abstract

Objective: The present study tests the hypothesis that changes in the glucose sensitivity of lateral hypothalamus (LH) hypocretin/orexin glucose-inhibited (GI) neurons following weight loss leads to glutamate plasticity on ventral tegmental area (VTA) dopamine neurons and drives food seeking behavior.

Methods: C57BL/6J mice were calorie restricted to a 15% body weight loss and maintained at that body weight for 1 week. The glucose sensitivity of LH hypocretin/orexin GI and VTA dopamine neurons was measured using whole cell patch clamp recordings in brain slices. Food seeking behavior was assessed using conditioned place preference (CPP).

Results: 1-week maintenance of calorie restricted 15% body weight loss reduced glucose inhibition of hypocretin/orexin GI neurons resulting in increased neuronal activation with reduced glycemia. The effect of decreased glucose on hypocretin/orexin GI neuronal activation was blocked by pertussis toxin (inhibitor of G-protein coupled receptor subunit Gα_i/o) and Rp-cAMP (inhibitor of protein kinase A, PKA). This suggests that glucose sensitivity is mediated by the Gα_i/o-adenylyl cyclase-cAMP-PKA signaling pathway. The excitatory effect of the hunger hormone, ghrelin, on hcrt/ox neurons was also blocked by Rp-cAMP suggesting that hormonal signals of metabolic status may converge on the glucose sensing pathway. Food restriction and weight loss increased glutamate synaptic strength (indexed by increased AMPA/NMDA receptor current ratio) on VTA dopamine neurons and the motivation to seek food (indexed by CPP). Chemogenetic inhibition of hypocretin/orexin neurons during caloric restriction and weight loss prevented these changes in glutamate plasticity and food seeking behavior.

Conclusions: We hypothesize that this change in the glucose sensitivity of hypocretin/orexin GI neurons may drive, in part, food seeking behavior following caloric restriction.

Keywords: Conditioned-place preference; Dopamine; Electrophysiology; Ghrelin; Glutamate plasticity; Protein kinase A.

Figure 1

Glucose sensing in hcrt/ox-GI neurons is mediated by protein kinase A (PKA). A. Representative whole cell voltage clamp recording of a hcrt/ox-GI neuron. When extracellular glucose decreased from 2.5 to 0.1 mM, this neuron reversibly depolarized its membrane potential (MP) and increased whole cell resistance (R) and action potential frequency. R was calculated by measuring the voltage response (downward deflection, expanded in inset to the right) to a constant hyperpolarizing current pulse (Ohm's Law: voltage (V) = current (I) x R). B–C. Percent change in MP and R in response to decreased glucose from 2.5 to 0.1 in the presence and absence (control) of: (B) the G-protein coupled receptor blocker, pertussis toxin (PTX, 500 ng/ml); brain slices were incubated in PTX for 2 h prior to recording (control, n/N (# neurons/#mice) = 6/4; PTX, n/N = 11/11, Independent Student's t-test) and (C) the PKA inhibitor, Rp-cAMP (10 μM; Rp) added to the bath solution (n/N = 5/5, Paired Student's t-test).∗∗∗P < 0.001, ∗∗∗∗P < 0.0001. D. Representative voltage responses to decreased glucose in the presence and absence of Rp-cAMP (gray: 2.5, black 0.1 mM glucose). Data represented as mean ± SEM.

Figure 2

Calorie restriction decreases glucose inhibition of hcrt/ox-GI neurons due, in part, to ghrelin action on PKA. A. Blood glucose levels (left) were measured in a subset of ad lib fed and calorie restricted (CR) mice at the start of the study (time 0), upon reaching ∼15% weight loss (WL) and after 1-week WL-maintenance (WL-M). Plasma glucose levels were lower during WL and WL-M (Repeated measures 2-way ANOVA showed an interaction between diet status and time (P < 0.0001, F (2,36) = 17.26); Sisak's multiple comparison ∗P < 0.05 CR compared to time 0 within their group, #P < 0.01 CR compared to the same time point for the ad lib fed group (n = 10/group). Plasma ghrelin levels (right) were significantly higher in these CR mice at the end of WL-M (Independent Student's t-test ∗∗∗P < 0.001, t = 4.074, df = 15, n = 7 ad lib, n = 10 CR) and plasma leptin levels (far right) were significantly lower (Independent Student's t-test ∗∗∗P < 0.0001, t = 5.59, df = 15, n = 8 ad lib, n = 9. CR). B. Top: Percent change in MP and R of hcrt/ox-GI neurons in response to decreased glucose from 2.5 to 0.7 in ad lib fed and CR mice (Independent Student's t-test, ∗P < 0.05, ∗∗P < 0.01, ad lib n/N = 6/5, CR n/N = 5/5). Bottom: Representative voltage responses of hcrt/ox-GI neurons to decreased glucose in ad lib and CR mice (gray: 2.5, black 0.1 mM glucose). C. Top: Percent change in MP and R of hcrt/ox neurons in response to the addition of ghrelin (10 nM) to a constant glucose level of 2.5 mM in the presence and absence of the PKA inhibitor Rp-cAMP (10 μM Rp, independent Student's t-test, ∗∗∗∗P < 0.0001, control n/N = 5/5, Rp-cAMP n/N = 6/5). Bottom: Representative voltage responses to ghrelin addition to 2.5 mM glucose in control and Rp-cAMP (gray: 2.5, black 0.1 mM glucose). Data represented as mean ± SEM.

Figure 3

Calorie restriction increases glutamatergic signaling onto VTA dopamine neurons and conditioned place preference (CPP) for a food reward. A. Calorie Restriction significantly increased α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA; left) but not N-methyl-d-aspartate (NMDA, right) receptor current amplitude. B. Left: Calorie Restriction significantly increased the AMPA/NMDA receptor current ratio. Independent Student's t-test, ad lib n/N = 8/5, calorie restricted (CR) n/N = 8/4, ∗∗P < 0.01, ∗∗∗P < 0.001. Right: Representative voltage clamp recordings of total (black) and NMDA receptor (grey) currents at a holding potential of 40 mV. NMDA currents were recorded in the presence of the AMPA inhibitor cyanquixaline (CNQX, 10 μM). AMPA currents were obtained by subtracting NMDA from total currents. C. Mean time spent on the food reward associated side of a CPP box for ad lib fed and CR mice before (pre-test) and after (test) conditioning for the food reward. Both ad lib fed and CR mice spent more time on the food reward associated side; however, CR mice spent significantly more time than did ad lib fed mice. Repeated measures 2-way ANOVA, Sidak's multiple comparison test; interaction between diet state and conditioning (F (1, 14) = 4.815, p < 0.05). Different letters indicate that groups are statistically different from each other (P < 0.05). Data represented as mean ± SEM.

Figure 4

Orexin mediates the effect of caloric restriction on glutamate currents on VTA dopamine neurons. A. There were no significant differences in AMPA or NMDA receptor current amplitude on VTA dopamine neurons from calorie restricted (CR) mice injected with control (pAAV-hSyn-DIO-mCherry) or inhibitory (pAAV-hSyn-DIO-hM4D(Gi)-mCherry) virus and given C21 (10 mg/kg) or vehicle (distilled water) in their drinking water (Two-way ANOVA followed by Tukey's post-hoc test, P > 0.05). B. The AMPA/NMDA receptor current ratio was significantly reduced in CR mice injected with the inhibitory DREADD and provided C21. Different letters indicate that groups are statistically different from each other (P < 0.05) (Two-way ANOVA followed by Tukey's post-hoc test, interaction between AAV and treatment ((F (1, 29) = 6.023, p < 0.05). Control-vehicle n/N = 8/4, control-C21 n/N = 9/5, Inhibitory DREADD-vehicle n/N = 8/5, Inhibitory DREADD-C21 n/N = 8/4. C. Representative voltage clamp recordings of total (black) and NMDA receptor (grey) currents at a holding potential of 40 mV. NMDA currents were recorded in the presence of the AMPA inhibitor cyanquixaline (CNQX, 10 μM). AMPA currents were obtained by subtracting NMDA from total currents. D. AMPA/NMDA receptor current ratio in VTA dopamine neurons from CR mice injected with the inhibitory DREADD and provided C21 in their drinking water in the presence (C21+, n/N = 8/4) or absence (C21-, n/N = 9/4) of C21 in the recording bath. There were no significant differences between the groups (Independent Student's t-test). Data represent mean ± SEM.

Figure 5

Orexin mediates the effect of calorie restriction on CPP. A. When compared before (pretest) and after (test) conditioning for a food reward, mean time spent on food associated side (sec) was significantly higher in all groups except the hM4Di group treated with C21 (Paired Student's t-test pre-test vs test for each group, ∗P < 0.05, nsd: not significantly different). B. Mean time spent on the food associated side. Two-way ANOVA showed an interaction between C21 treatment and AAV ((F (1, 29) = 13.04, p < 0.01). Tukey's post-hoc test showed a significant decrease in the mean time spent on the food associated in hM4Di mice treated with C21 (n = 9) compared to control mice treated with vehicle (n = 8) or C21 (n = 8) and hM4Di mice treated with vehicle (n = 8). C. Mean percentage time on the conditioned side on the testing day. Two-way ANOVA showed a significant interaction between C21 treatment and AAV ((F (1, 29) = 14.77, p < 0.01). Tukey's post-hoc test showed a significant decrease in the mean percentage time spent on the food associated side in hM4Di mice treated with C21 compared to control mice treated with vehicle or C21 and hM4Di mice treated with vehicle. The mean percentage time in C21 treated mice injected with hM4Di in the food associated side was close to 50%, suggesting a failure to attain place conditioning to food-reward. Different letters indicate that groups are statistically different from each other (P < 0.05).