Hedonic eating is controlled by dopamine neurons that oppose GLP-1R satiety

Abstract

Hedonic eating is defined as food consumption driven by palatability without physiological need. However, neural control of palatable food intake is poorly understood. We discovered that hedonic eating is controlled by a neural pathway from the peri-locus ceruleus to the ventral tegmental area (VTA). Using photometry-calibrated optogenetics, we found that VTA dopamine (VTA^DA) neurons encode palatability to bidirectionally regulate hedonic food consumption. VTA^DA neuron responsiveness was suppressed during food consumption by semaglutide, a glucagon-like peptide receptor 1 (GLP-1R) agonist used as an antiobesity drug. Mice recovered palatable food appetite and VTA^DA neuron activity during repeated semaglutide treatment, which was reversed by consumption-triggered VTA^DA neuron inhibition. Thus, hedonic food intake activates VTA^DA neurons, which sustain further consumption, a mechanism that opposes appetite reduction by semaglutide.

Fig. 1.. Closed-loop photoinhibition of periLC^VGLUT2 axon projections to the VTA prolongs consumption.

(A) Feeding bout analysis to compare consumption of higher and lower palatability food available in alternating 2-minute blocks (teal and grey blocks, respectively). E: Ensure. (B-G) Higher palatability 100% Ensure leads to more food intake (B, E), longer bout duration (C, F), and a similar number of bouts initiated (D, G) compared to lower palatability diluted Ensure (B-D) (n = 11 mice) or 100% Ensure adulterated with 3 mM quinine (E-G) (n = 7 mice). (H) Cre-dependent expression of ArchT-EGFP in periLC neurons in Vglut2-IRES-Cre mice and EGFP fluorescence in downstream brain regions. (I) Summary of periLC^VGLUT2 neuron axon projection targets. (J) Schematic of periLC^VGLUT2 axon projection photoinhibition at BNST, LHA, VTA, PCRt during alternating ON and OFF blocks of lick-triggered axon photoinhibition while consuming lower palatability food. (K) ArchT-EGFP fluorescence and highlighted bilateral optical fiber tracts in the BNST, LHA, VTA, and PCRt. Scale bar, 500 μm. (L) Mean licks per block for food in alternating ON and OFF blocks during lick-triggered photoinhibition of periLC^VGLUT2 axons projections to BNST, LHA, VTA, PCRt (paired t-test, n = 3,6,9,4 mice). (M-P) Feeding bout analysis of laser-ON and laser-OFF blocks for lick-triggered photoinhibition of periLC^VGLUT2 axons in the BNST, LHA, VTA, and PCRt (KS-test and paired t-test, n = 3,6,9,4 mice). (Q) Schematic of lick-triggered photoactivation of periLC^VGLUT2→VTA axon projections during feeding. (R) EGFP expression in the periLC and fiber tracts in the VTA. Scale bar, 500 μm. (S-V) Lick-triggered photoactivation of periLC^VGLUT2 axon projections to the VTA suppressed food intake (S), reduced bout duration (T-U), and reduced bout initiation (V) (KS-test and paired t-test, n = 4). Data are represented as mean ± SEM. ns p>0.05, *p<0.05, **p<0.01, ***p<0.001. Statistical details are in Table S1.

Fig. 2.. periLC^VGLUT2 neurons suppress VTA^DA neuron activity and NAc dopamine.

(A) Photometry and strategy for expression of Chrimson in periLC^VGLUT2 neurons and GCaMP8m in VTA^DA neurons in Vglut2-Flpo; Slc6a3-IRES-Cre mice. (B, C) Expression of Chrimson-tdTomato (red) and anti-tyrosine hydroxylase immunoreactivity (TH, yellow) in periLC and LC, respectively (B). GCaMP8m (green, left), and its overlay (right) with TH (magenta) in VTA^DA neurons, and optical fiber track (grey) (C). Arrows indicate GCaMP8m+/TH+ neurons. Scale bars, 200 and 100 μm. (D) Expression of Chrimson (red) with GCaMP8m (green, top) or VGAT (white, bottom) in VTA. Scale bar, 100 μm. (E) VTA^DA neuron calcium dynamics during photostimulation (PS) of periLC^VGLUT2→VTA axons. (F) AUC from VTA^DA GCaMP8m decreases during PS (paired t-test, n = 4 mice). (G-I) GCaMP8m mean responses during consumption of 100% Ensure (G) and 100% Ensure with lick-contingent PS of periLC^VGLUT2→VTA axons (H) show decreased bout duration-normalized GCAMP8m response (I, paired t-test, n = 4 mice). (J-N) Lick-contingent PS of periLC^VGLUT2→VTA axons decreases consumption (J, K), and bout duration (L, M) but not bout number (N) during laser-ON blocks (negative binomial generalized linear mixed models, KS-test, and paired t-test, n = 4). (O) Photometry setup and viral transduction strategy for Cre-dependent expression of Chrimson in periLC^VGLUT2 neurons and expression of GRAB-DA2m in NAc (Vglut2-IRES-Cre mice). (P, Q) Expression of Chrimson (red) adjacent to TH (yellow) in the periLC and LC, respectively (P) and GRAB-DA2m (green) in NAc (Q), Scale bars, 200 μm. (R) Z-scored GRAB-DA2m fluorescence during PS of periLC^VGLUT2 neurons. (S) AUC of GRAB-DA2m fluorescence decreases during PS. (T-V) GRAB-DA mean responses during consumption of 100% Ensure (T) and 100% Ensure with lick-contingent PS of periLC^VGLUT2 neurons (U) show decreased bout duration-normalized NAc dopamine (V, paired t-test, n = 5 mice). (W) Circuit diagram. Data are represented as mean ± SEM. ns p>0.05, *p<0.05, **p<0.01, ***p<0.001. Statistical details are provided in Table S1.

Fig. 3.. VTA^DA neuron activity and NAc dopamine are dependent on consumption duration and palatability.

(A) Photometry of VTA^DA neuron calcium dynamics during consumption of 20% or 100% Ensure in separate sessions (Slc6a3-IRES-Cre mice). (B-E) GCaMP8s activity during consumption of 20% (B, D) and 100% Ensure (C, E). (D, E) GCaMP8s mean responses (blue) during consumption of 20% Ensure and 100% Ensure. The magenta line is the variable-length time mean response, which takes variable bout duration into account (see Methods). (F-G) Regression of GCaMP8s AUC with bout duration for consumption of 20% Ensure (F) and 100% Ensure (G) (n = 13 mice). (H-I) GCaMP8s AUC (H) slope (I) (paired t-test, n = 13 mice). (J) Setup for consumption of 20% or 100% Ensure during the same session (Slc6a3-IRES-Cre mice). (K-N) GCaMP8s during consumption of 20% (K, M) and 100% Ensure (L, N). (M, N) GCaMP8s mean responses (blue) and variable-length time mean response (magenta) during consumption of 20% Ensure and 100% Ensure. (O-P) Regression of GCaMP8s AUC with bout duration for 20% Ensure (O) and 100% of Ensure (P) (n = 8 mice). (Q-R) GCaMP8s AUC (Q) and slope (R) (paired t-test, n = 8 mice). (S) Photometry setup to measure NAc dopamine during consumption of 20% or 100% Ensure in separate sessions. (T-W) GRAB-DA responses during consumption of 20% (T, V) and 100% Ensure (U, W). (V, W) GRAB-DA mean responses (blue) and variable-length time mean response (magenta) during consumption of 20% Ensure and 100% Ensure. (X-Y) Regression of GRAB-DA AUC with bout duration during consumption of 20% Ensure (X) and 100% Ensure (Y) (n = 9 mice). (Z-AA) GRAB-DA AUC (Z) and slope (AA) (paired t-test, n = 9 mice). (AB) Photometry for GRAB-DA during consumption of 20% or 100% Ensure in the same session. (AC-AF) GRAB-DA responses during consumption of 20% (AC, AE) and 100% Ensure (AD, AF). (AE, AF) GRAB-DA mean responses (blue) and variable-length time mean response (magenta) during consumption of 20% Ensure and 100% Ensure. (AG-AH) Regression of GRAB-DA AUC with bout duration during consumption of 20% Ensure (AG) and 100% of Ensure (AH) (n = 13 mice). (AI-AJ) GRAB-DA AUC (AI) and slope (AJ) (paired t-test, n = 13 mice). Data are represented as mean ± SEM. ns p>0.05, **p<0.01, ***p<0.001. Statistical details are in Table S1.

Fig. 4.. VTA^DA neurons promote food consumption duration and palatability.

(A) Photometry-calibrated photostimulation (PS) setup. E: Ensure. (B) Expression of GCaMP8s (green, top) and overlay with anti-TH (red, bottom) within VTA. Arrows indicate GCaMP8s+ TH+ neurons. Scale bar, 100 μm. (C) Co-expression among GCaMP8s (left) and anti-TH cells (right) within the VTA. (D, E) Recorded VTA^DA neuron activity during 20% Ensure (D, left), 100% Ensure (D, right), and 20% Ensure with photometry-calibrated VTA^DA neuron PS (E). Vertical line: First lick of a bout. (F-I) GCaMP8s responses during consumption of 20% Ensure with photometry-calibrated VTA^DA neuron PS in ON-blocks (F, H) and 20% Ensure in OFF-blocks (G, I). (H, I) GCaMP8s mean responses (blue) and mean response within a bout (magenta) during consumption of 20% Ensure with photometry-calibrated VTA^DA neuron PS (H) and 20% Ensure (I). (J-K) Regression of GCaMP8s AUC with bout duration during consumption of 20% Ensure with photometry-calibrated VTA^DA neuron PS (J) and 20% Ensure (K) (n = 13 mice). (L-M) GCaMP8s AUC (L) and slope (M) for photometry-calibrated VTA^DA neuron PS (paired t-test, n = 13 mice). (N-S) Lick-contingent photometry-calibrated PS of VTA^DA neurons increases consumption (N), and bout duration but not bout number during laser-on blocks (O-R), which is apparent after 3-s of PS onset (S) (negative binomial generalized linear mixed model, KS-test and paired t-test, n = 13 mice). (T-Y) Noncontingent PS does not affect licking behaviors (T), nor bout duration and bout number (U-Y) (negative binomial generalized linear mixed model, KS-test and paired t-test, n = 8 mice). (Z) Dopamine-calibrated photometry and PS setup. (AA-AB) NAc dopamine responses during 20% Ensure (AA, left), 100% Ensure (AA, right), and 20% Ensure with photometry-calibrated VTA^DA neuron PS (AB). (AC-AD) Lick-contingent photometry-calibrated PS of VTA^DA neurons increases bout duration (AC) but not bout number (AD) during laser-ON blocks (paired t-test, n = 5 mice). Data are represented as mean ± SEM. ns p>0.05, **p<0.01. ***p<0.001. Statistical details are in Table S1.

Fig. 5.. VTA^DA neurons are necessary for prolonged food consumption duration.

(A) Left panel, setup for photoinhibition of VTA^DA neurons using AAV-CAG-FLEX-JAWS-GFP in mice consuming 100% Ensure (Slc6a3-IRES-Cre mice). Right panel, the viral expression of FLEX-JAWS within VTA (green), the fiber tract, and the overlayed histological images with Anti-TH (red). Scale bar, 200 μm. (B) Schematic for lick-contingent or noncontingent photoinhibition (2 min ON or OFF blocks). (C-F) Lick-contingent photoinhibition of VTA^DA neurons reduced licking behaviors (C) and bout duration but not bout number (D-E), which is apparent after 5-s of PS onset (F) during laser-ON blocks (negative binomial generalized linear mixed model, KS-test and paired t-test, n = 6 mice). (G-J) Noncontingent photoinhibition did not significantly affect licking behaviors (G), bout duration, or bout number (H-J) (negative binomial generalized linear mixed model, KS-test and paired t-test, n = 6 mice). Data are represented as mean ± SEM. ns p>0.05, *p<0.05, ***p<0.001. Statistical details are in Table S1.

Fig. 6.. Semaglutide modulates VTA^DA neurons to shorten consumption bout duration.

(A) Experimental design to test the effects of semaglutide (doses in mg/kg) on food consumption. (B-F) Behavioral summary of body weight (B), homecage chow food intake (C), 2-hour Ensure intake (D), bout duration (E), and bout numbers (F) during semaglutide or PBS injection (n = 8 mice). (G) Photometry setup for recording VTA^DA neuron activity during 100% Ensure intake with semaglutide and PBS injection (Slc6a3-IRES-Cre mice). (H-L) Behavioral summary of body weight (H), homecage food intake (I), 1-hour Ensure intake (J), bout duration (K), and bout numbers (L) during Semaglutide and PBS injection (n = 8 mice). (M) GCaMP8s response of VTA^DA neuron activity to 100% Ensure from day 0 to day 5 during Semaglutide injection (n = 8 mice). (N) Comparison of GCaMP responses to 100% Ensure in VTA^DA neurons on day 0 to day 5 with PBS and Semaglutide injections (rmANOVA, n = 8 mice). Data are represented as mean ± SEM. ns p>0.05, *p<0.05, **p<0.01, ***p<0.001. Statistical details are in Table S1.

Fig. 7.. VTA^DA neurons bidirectionally control palatable food consumption with semaglutide treatment.

(A) Dual photometry and photostimulation for recording and activating VTA^DA neuron activity during food consumption with Semaglutide injection. (B) GCaMP response of VTA^DA neuron activity to 100% Ensure during photostimulation of VTA^DA neurons following semaglutide injection on Day 1. (C-E) Lick-contingent photostimulation of VTA^DA neurons increased lick numbers (C) and bout duration (D) but not bout number (E) in laser-ON blocks (KS-test and paired t-test, n = 10 mice). (F) Experimental design for photoinhibition of VTA^DA neurons during day 4 to day 6 with the highest dose of semaglutide. (G-H) Body weight (G) and home cage chow food intake (H) during semaglutide treatment. (I-N) Reduced Ensure intake (I, J), and bout duration (K, L) but not bout numbers (M, N) during the laser-ON blocks compared to laser-OFF blocks for Days 4–6 (n = 10 mice). Days 1–3 are analyzed to show Ensure intake, bout duration, and bout number across the same alternating 2-min blocks (B1, B2) in the absence of photoinhibition. Data are represented as mean ± SEM. ns p>0.05, *p<0.05, **p<0.01, ***p<0.001. Statistical details are in Table S1.

Fig. 8.. Summary of VTA^DA neuron activity during feeding behaviors.

VTA^DA neuron activity in different phases of food intake.