Distinct Signaling by Ventral Tegmental Area Glutamate, GABA, and Combinatorial Glutamate-GABA Neurons in Motivated Behavior

Abstract

Ventral tegmental area (VTA) neurons play roles in reward and aversion. We recently discovered that the VTA has neurons that co-transmit glutamate and GABA (glutamate-GABA co-transmitting neurons), transmit glutamate without GABA (glutamate-transmitting neurons), or transmit GABA without glutamate (GABA-transmitting neurons). However, the functions of these VTA cell types in motivated behavior are unclear. To identify the functions of these VTA cell types, we combine recombinase mouse lines with INTRSECT2.0 vectors to selectively target these neurons. We find that VTA cell types have unique signaling patterns for reward, aversion, and learned cues. Whereas VTA glutamate-transmitting neurons signal cues predicting reward, VTA GABA-transmitting neurons signal cues predicting the absence of reward, and glutamate-GABA co-transmitting neurons signal rewarding and aversive outcomes without signaling learned cues related to those outcomes. Thus, we demonstrate that genetically defined subclasses of VTA glutamate and GABA neurons signal different aspects of motivated behavior.

Keywords: GABA; aversion; glutamate; motivation; reward; ventral tegmental area.

Figure 1.. Selective Targeting of VTA VGluT2⁺ VGaT⁺, VGluT2⁺ VGaT⁻, and VGluT2⁻ VGaT⁺ Neurons

(A and B) Schematic of crossing of vglut2-Cre mice with vgat-Flp mice to generate vglut2-Cre/vgat-Flp mice and intra-VTA injections of INTRSECT2.0 AAV-C_ON/F_ON-mCherry (targeting VGluT2⁺ VGaT⁺ neurons), AAV-C_ON/F_OFF-eYFP (targeting VGluT2⁺ VGaT⁻ neurons), or AAV-C_OFF/F_ON-BFP vectors (targeting VGluT2⁻ VGaT⁺ neurons). INTRSECT vectors split the fluorophore sequence in half (N or C terminus) between two independent lox sites (boxed Cre-dependent triangles are loxN and lox2722) and FRT sites (boxed Flp-dependent triangles are F3 and F5). Lox and FRT sequences between N and C termini are within introns, which are spliced out for recombination of the entire fluorophore sequence. (C) Co-expression of VGluT2 mRNA and VGaT mRNA in VTA VGluT2⁺ VGaT⁺ mCherry neurons. (D) VGluT2 mRNA without VGaT in VTA VGluT2⁺ VGaT⁻ eYFP neurons. (E) VGaT mRNA without VGluT2 in VTA VGluT2⁻ VGaT⁺ BFP neurons. (F–H) Detection of VGluT2 or VGaT mRNAs within the subpopulations of neurons co-expressing mCherry (F), eYFP (G), or BFP (H). Black dots are individual mice (N = 3/group).

Figure 2.. Projections of VTA VGluT2⁺ VGaT⁺, VGluT2⁺ VGaT⁻, and VGluT2⁻ VGaT⁺ Neurons

(A) Schematic of injection of a cocktail of INTRSECT2.0 vectors (AAV-C_ON/F_ON-mCherry, AAV-C_ON/F_OFF-eYFP, and AAV-C_OFF/F_ON-BFP) into the VTA of vglut2-cre/vgat-Flp mice. (B) Detection of mCherry (targeting VGluT2⁺ VGaT⁺ neurons), eYFP (targeting VGluT2⁺ VGaT⁻ neurons), BFP (targeting VGluT2⁻ VGaT⁺ neurons), and TH immunoreactivity in VTA at low magnification. (C) At higher magnification, visualization of VTA mCherry, eYFP, BFP, or TH cell bodies. (D) Lateral habenula (LHb) with dense mCherry fibers and less dense eYFP or BFP fibers. (E) Nucleus accumbens shell (nAcc shell) with dense eYFP fibers and less dense mCherry or BFP fibers. (F) Bed nucleus of the stria terminalis (BNST) with BFP fibers and less dense mCherry or eYFP fibers.

Figure 3.. Vesicular Accumulation and Release of Glutamate or GABA by VTA VGluT2⁺ VGaT⁺, VGluT2⁺ VGaT⁻, and VGluT2⁻ VGaT⁺ Neurons

(A) AAV-C_ON/F_ON-ChR2-eYFP injection in VTA of vglut2-cre/vgat-Flp mice. (B) LHb detection of eYFP, VGluT2, and VGaT, showing VGluT2 and VGaT expression in eYFP-positive axon terminals (arrows). (C) LHb co-expression of VGluT2 and VGaT in 97.04% of eYFP axon terminals (n = 3). (D) AAV-C_ON/Flp_OFF-ChR2-eYFP injection in VTA of vglut2-cre/vgat-Flp mice. (E) nAcc shell detection of eYFP, VGluT2, and VGaT, showing expression of VGluT2 in eYFP axon terminals, but not VGaT (arrows). (F) nAcc shell expression of VGluT2 without VGaT in 95.04% of eYFP axon terminals (n = 3). (G) AAV-C_OFF/F_ON-ChR2-eYFP injection in VTA of vglut2-cre/vgat-Flp mice YFP. (H) BNST detection of eYFP, VGluT2, and VGaT, showing expression of VGaT in eYFP axon terminals, but not VGluT2 (arrows). (I) BNST expression of VGaT without VGluT2 in 96.41% of eYFP axon terminals (n = 3). (J–R) Release of glutamate or GABA by VTA axon terminals. (J) AAV-C_ON/F_ON-ChR2-eYFP injection in VTA and LHb recordings. (K and L) LHb photostimulation (blue line) evoked excitatory postsynaptic currents (EPSCs) at −60 mV and inhibitory postsynaptic currents (IPSCs) at 0 mV. Currents were inhibited by TTX(at −60 mV: −38.9 ± 9.7 pA baseline and −7.9 ± 1.8 pA after TTX; at 0 mV: 162.5 ± 50.4 pA baseline and 12.6 ± 3.9 pA after TTX). Currents were restored by 4-AP (at −60mV: −38.9 ± 9.7 pA baseline and −45.8 ± 13.9 pA after 4-AP; at 0 mV: 62.5 ± 50.4 pA baseline and 178.2 ± 26.8 pA after 4 AP). EPSCs were blocked by CNQX (13.0 ± 2.9 pA) and IPSCs by bicuculline (8.2 ± 1.4 pA; −60 mV currents F_(4,29) = 13.09; ***p ^< 0.0001; repeated-measures ANOVA; Dunnett post hoc test n = 6 cells of 4 mice; 0 mV currents F_(4,29) = 13.89; ****p ^< 0.0001; repeated-measures ANOVA; Dunnett post hoc test n = 6 cells of 4 mice; *p ^< 0.05; **p ^< 0.01; ***p ^< 0.01). (M) AAV-C_ON/F_OFF-ChR2-eYFP injection in VTA and nAcc recordings. (N and O) nAcc photostimulation evoked EPSCs, which were inhibited by TTX (−44.1 ± 5.6 pA baseline and −7.4 ± 1.2 pA after TTX), restored by 4-AP (−44.1 ± 5.6 pA baseline and 34.7 ± 5.0 pA after 4-AP), and blocked by CNQX (4.7 ± 0.7 pA). (P) AAV-C_OFF/F_ON-ChR2-eYFP injection in VTA and BNST recordings. (Q and R) BNST photostimulation evoked IPSCs, which were inhibited by TTX (91.6 ± 29.0 pA baseline; 6.5 ± 2.5 pA after TTX), restored by 4-AP (182.0 ± 67.9 pA), and blocked by bicuculline (8.3 ± 4.0 pA). Data are mean ± SEM.

Figure 4.. Response to Conditioned Reward by VTA VGluT2⁺ VGaT⁺, VGluT2⁺ VGaT⁻, VGluT2⁻ VGaT⁺, and TH⁺ Neurons

(A) Schematic of injection of AAV-C_ON/F_ON-GCAMP6m,AAV-C_ON/Flp_OFF-GCAMP6m, or AAV-C_OFF/F_ON-GCAMP6m into VTA of vglut2-Cre/vgat-Flp mice or AAV-DIO-GCaMP6m into the VTA of th-Cre mice. optic fiber placement over GCAMP-positive cells in the VTA is shown. (B) Diagram of calcium imaging system. (C) Sucrose reward conditioning and testing procedures. Mice learned to respond more to the CS+ over training (p < 0.05) and to discriminate between CS+ and CS− (p < 0.0001; N = 6–10 per group; two-way ANOVA; session × cue type interaction; ANOVA; F_(1,28) = 5.69; p < 0.024). *p < 0.05 (D–G) Observed Ca²⁺ responses in the VTA during sucrose reward conditioning (left) and population average Ca²⁺ signals (right). BL, baseline. Data are mean ± SEM. Comparisons between event-related activity are as follows: *p < 0.05, **p < 0.01, and ***p < 0.0001. (D) VGluT2⁺ VGaT⁺ neurons showing lack of changes in Ca²⁺ signal by the CS+ or CS− but increased signal by sucrose delivery (p < 0.001). (E) VGluT2⁺ VGaT⁻ neuron showing increases in Ca²⁺ signal by the CS+ (p < 0.01) and by sucrose delivery (p < 0.001). (F) VGluT2⁻ VGaT⁺ neurons showing moderate increase in Ca²⁺ signal by the CS+ (p = 0.064), sucrose delivery (p < 0.01), and by the CS− (p = 0.041). (G) TH⁺ neurons showing increases in Ca2⁺ signal by the CS+ (p < 0.001) and by sucrose delivery (p < 0.01; N = 6–10 per group; mixed ANOVA; group × cue type × event interaction; F_(12,108) = 4.802; p < 0.0001; Sidak-correction for pos-hoc comparisons).

Figure 5.. Response to Errors in the Prediction of Reward by VTA VGluT2⁺ VGaT⁺, VGluT2⁺ VGaT⁻, VGluT2⁻ VGaT⁺, and TH⁺ Neurons

(A) Schematic of injection of AAV-C_ON/F_ON-GCAMP, AAV-C_ON/Flp_OFF-GCAMP6m, or AAV-C_OFF/F_ON-GCAMP6m into VTA of vglut2-Cre/vgat-Flp mice or AAV-DIO-GCaMP6m into the VTA of th-Cre mice. Optic fiber placement over GCAMP-positive cells in the VTA is shown. (B) Testing procedures in which VTA cell responses were recorded in response to CS+ and sucrose delivery (expected reward), CS+ and sucrose omission (reward-omission, error), or CS+ omission and sucrose delivery (unexpected reward). (C–F) Observed population responses during sucrose reward conditioning (left) and population average Ca²⁺ signals for each event (right). Data are mean ± SEM. Comparisons between event-related activity are as follows: *p < 0.05 and **p < 0.01. Comparisons to baseline activity are as follows: #p < 0.05; ##p < 0.01; and ###p < 0.001. (C) VGluT2⁺ VGaT⁺ neurons showing higher Ca²⁺ signal in response to expected sucrose versus its omission and even higher Ca²⁺ activity in response to unexpected sucrose (reward versus omission trials, p < 0.01; unexpected reward versus omission trials, p < 0.05). (D) VGluT2⁺ VGaT⁻ neurons showing similar increases in Ca²⁺ signals in response to expected sucrose, sucrose omission (p < 0.01), or unexpected sucrose (p < 0.01). (E) VGluT2⁻ VGaT⁺ neurons showing similar increase in Ca²⁺ signaling in response to expected sucrose, sucrose omission (p < 0.05), or unexpected sucrose (p < 0.001). (F) TH⁺ neurons higher Ca²⁺ signaling in response to expected or unexpected sucrose than when sucrose was omitted (p < 0.05).

Figure 6.. Response to Aversive Stimuli by VTA VGluT2⁺ VGaT⁺, VGluT2⁺ VGaT⁻, VGluT2⁻ VGaT⁺, and TH⁺ Neurons

(A) Schematic of injection of AAV-C_ON/F_ON-GCAMP, AAV-C_ON/Flp_OFF-GCAMP6m, or AAV-C_OFF/F_ON-GCAMP6m into VTA of vglut2-Cre/vgat-Flp mice or AAV-DIO-GCaMP6m into the VTA of th-Cre mice. Optic fiber placement over GCAMP-positive cells in the VTA is shown. (B) Schematic of fear-conditioning training and cue-test sessions. (C) Heatmaps of Ca²⁺ signal over successive trials of fear-conditioning (during cue and footshock presentations). (D) Cell population responses to cue and shock in conditioning training showing increases in Ca²⁺signal in VGluT2⁺ VGaT⁺, VGluT2⁺ VGaT⁻, VGluT2⁻ VGaT⁺, and TH⁺ neurons in response to footshock (VGluT2⁺ VGaT⁺, p < 0.001; VGluT2⁺ VGaT⁻, p< 0.001; VGluT2⁻ VGaT⁺, p < 0.001; TH⁺, p< 0.05) and in response to the cue predicting the shock for VGluT2⁺ VGaT⁻ (p < 0.05) and VGluT2⁻ VGaT⁺ cell populations (p < 0.05). Data are mean ± SEM. Comparisons between event-related activity are as follows: *p < 0.05, **p < 0.01, ***p < 0.001. (E) Cell population average Ca²⁺ signals show increases in Ca²⁺ during the cue test for VGluT2⁺ VGaT⁻ (p< 0.05) and VGluT2⁻ VGaT⁺ neurons (p< 0.05), demonstrating a memory for the shock-paired cue. Data are mean ± SEM. Comparisons between event-related activity are as follows: *p < 0.05, **p < 0.01, ***p < 0.001.