Ethanol-seeking behavior is expressed directly through an extended amygdala to midbrain neural circuit

Abstract

Abstinent alcohol-dependent individuals experience an enduring sensitivity to cue-induced craving and relapse to drinking. There is considerable evidence indicating that structures within the midbrain and extended amygdala are involved in this process. Individually, the ventral tegmental area (VTA) and the bed nucleus of the stria terminalis (BNST) have been shown to modulate cue-induced ethanol-seeking behavior. It is hypothesized that cue-induced seeking is communicated through a direct projection from the BNST to VTA. In the current experiments, an intersectional viral strategy was used in DBA/2J mice to selectively target and inhibit BNST projections to the VTA during a test of ethanol conditioned place preference (CPP). Inhibitory designer receptors exclusively activated by designer drugs (hM4Di DREADDs) were expressed in VTA-projecting BNST (BNST-VTA) cells by infusing a retrograde herpes-simplex virus encoding cre recombinase (HSV-Cre) into VTA and a cre-inducible adeno-associated virus encoding hM4Di (AAV-DIO-hM4Di) into BNST. Before testing the expression of preference, clozapine-N-oxide (CNO) was peripherally administered to activate hM4Di receptors and selectively inhibit these cells. Ethanol CPP expression was blocked by CNO-mediated inhibition of BNST-VTA cells. A follow-up study revealed this effect was specific to CNO activation of hM4Di as saline- and CNO-treated mice infused with a control vector (HSV-GFP) in place of HSV-Cre showed significant CPP. These findings establish a role for a direct BNST input to VTA in cue-induced ethanol-seeking behavior.

Keywords: BNST; CPP; DREADD; Ethanol; HSV; VTA.

Figure 1. Expression of hM4Di receptors in VTA-projecting BNST neurons using a retrograde intersectional strategy

(A) A dual-viral approach was used to drive expression of inhibitory designer receptors (hM4Di) in a distinct yet intermixed subpopulation of BNST neurons that project to VTA. A long-term retrograde HSV encoding cre recombinase was delivered into the VTA and a cre-dependent AAV-hM4Di was delivered into the BNST. (B) hM4Di-mCherry AAV vector design employing the double-floxed inverted open reading frame (DIO) strategy. Two pairs of heterotypic, antiparallel loxP-type recombination sites achieve cre-mediated hM4Di inversion and expression under the control of a human synapsin (hSyn) promoter. WPRE, woodchuck hepatitis virus post-transcriptional regulatory element; ITR, inverted terminal repeat, pA, human growth hormone polyadenylation site (C) Sagittal section showing intersection of retrogradely transported HSV-cre (VTA-projecting cre+ cells are pseudocolored magenta) and AAV expressing cre-dependent hM4Di (hM4Di+ BNST-VTA cells are pseudocolored cyan) 8 weeks after vector infusions. Robust expression of hM4Di is visible in soma (in BNST) and fibers (to VTA) of BNST-VTA neurons. HSV-cre transfected neurons projecting to VTA are visible throughout the brain, with the exception of VTA as retrograde HSV does not infect cell bodies at the site of injection. Note the absence of hM4Di in all HSV-Cre transfected cells outside the BNST. This illustrates that hM4Di is localized to VTA projecting neurons within the BNST only. D, dorsal; V, ventral; A, anterior; P, posterior; scale bar, 1 mm.

Figure 2. Ethanol-induced CPP expression is blocked by activation of hM4Di receptors selectively expressed in BNST-VTA cells

(A) Heterologous expression of cre-dependent hM4Di (red; visualized by IF detection of mCherry) is observed in VTA-projecting neurons (green; visualized by IF detection of EYFP) of dorsal and ventral BNST. Nuclei are counterstained blue with DAPI; ac, anterior commissure; scale bar, 200 µm. (B) High magnification image of region outlined in white box above illustrating HSV-mediated cre expression (green), DIO-hM4Di expression (red), and their overlap (orange-yellow), with nuclei stained blue. Scale bar, 50 µm. (C) Mean (+SEM) time spent on the grid floor (in s/min) during 30-min preference test. VTA-projecting BNST neurons were inhibited via CNO (10 mg/kg)-mediated stimulation of hM4Di. Inhibition of BNST-VTA signaling blocked the expression of ethanol-induced CPP. † p≤0.001 interaction between drug and conditioning subgroup (Grid+ vs. Grid−); * p<0.001 between conditioning subgroups; n=13–15/subgroup. (D) Mean percent time (±SEM) spent on the ethanol-paired floor in 5-min intervals across the 30-min preference test. BNST-VTA inhibition reduced place preference in a consistent manner across the test.

Figure 3. Ethanol CPP expression is not disrupted by CNO in mice expressing GFP in VTA-projecting BNST cells

(A) Expression of GFP in dBNST and vBNST 8 weeks after infusion of HSV-GFP into VTA and DIO-hM4Di into BNST. GFP+ cells (green) indicate VTA-projecting neurons. No hM4Di expression was visible in BNST. ac, anterior commissure; scale bar, 500 µm. (B) High magnification image from BNST of GFP (green; visualized by IF detection of GFP), hM4Di (absent; visualized by IF detection of mCherry), nuclei (DAPI) and all channels merged. Note that hM4Di is not expressed in the absence of cre; scale bar, 50 µm. (C) Mean (+SEM) time spent on the grid floor (in s/min) during 30-min preference test. CNO did not disrupt ethanol CPP expression in mice expressing GFP in BNST-VTA cells; n=11–12 / conditioning subgroup (Grid+, Grid). (D) There was no significant difference in percent time spent on the ethanol-paired floor between groups when analyzed in 5-min intervals across the 30-min preference test.