GLP-1 acts on habenular avoidance circuits to control nicotine intake

Abstract

Tobacco smokers titrate their nicotine intake to avoid its noxious effects, sensitivity to which may influence vulnerability to tobacco dependence, yet mechanisms of nicotine avoidance are poorly understood. Here we show that nicotine activates glucagon-like peptide-1 (GLP-1) neurons in the nucleus tractus solitarius (NTS). The antidiabetic drugs sitagliptin and exenatide, which inhibit GLP-1 breakdown and stimulate GLP-1 receptors, respectively, decreased nicotine intake in mice. Chemogenetic activation of GLP-1 neurons in NTS similarly decreased nicotine intake. Conversely, Glp1r knockout mice consumed greater quantities of nicotine than wild-type mice. Using optogenetic stimulation, we show that GLP-1 excites medial habenular (MHb) projections to the interpeduncular nucleus (IPN). Activation of GLP-1 receptors in the MHb-IPN circuit abolished nicotine reward and decreased nicotine intake, whereas their knockdown or pharmacological blockade increased intake. GLP-1 neurons may therefore serve as 'satiety sensors' for nicotine that stimulate habenular systems to promote nicotine avoidance before its aversive effects are encountered.

Figure 1. Nicotine activates medullary GLP-1 neurons

(a) Graphical representation of GLP-1-positive (GLP-1+) and tyrosine hydroxylase-positive (TH+) neurons in the cNTS (cc: central canal). (b) Representative micrographs of Fos (in red) and GLP-1+ or TH+ (in green) neurons in the cNTS and VLM after saline or nicotine (0.25–1.5 mg/kg) injection (insert: 20x magnification of marked area). (c) Relative neuronal recruitment by cell type in cNTS following nicotine challenge. Scale bar, 100μm.

Figure 2. GLP-1 regulates nicotine intake

(a) Graphical representation of mechanisms of GLP-1 receptor (GLP-1R) activation by sitagliptin and exendin-4 (Ex-4). (b) Graphical representation of intravenous nicotine self-administration procedure. (c) Mean (± s.e.m.) number of nicotine infusions earned by mice after Ex-4 (10 μg/kg) administration. ***P<0.001, paired t-test (n=10). (d) Mean (± s.e.m.) number of nicotine infusions earned by mice after sitagliptin administration. **P<0.01, paired t-test (n=12). (e) Nicotine dose-response curve of GLP-1R knockout (n=8) and wild-type (n=6) mice. Data points represent mean (± s.e.m.) number of infusions earned at each nicotine dose. Two-way repeated-measures (RM) ANOVA, Genotype: F_{(1, 12)}=6.29, *p<0.05; Nicotine: F_{(2, 24)}=9.49, p<0.001; Genotype × Nicotine: F_{(2, 24)}=0.18, NS. (f) Mean (± s.e.m.) number of food rewards earned during 1-hour operant training sessions between GLP-1R knockout (n=9) and wild-type (n=12) mice. Scale bar, 100 μm.

Figure 3. Chemogenetic activation of GLP-1 neurons mice decreases nicotine intake

(a) Graphical representation of NTS and the site of stereotaxic DREADD virus injections in Phox2b-Cre of Gcg-Cre mice. (b) Upper panel: DIO vector design and selective expression in Phox2b-Cre neurons. Lower panel: Representative micrograph showing selective expression of AAV-DIO-hM3Dq-mCitrine vector in GLP-1-immunopositive neurons in Phox2b-Cre mice. (c) Non-DIO vector design (upper panel) and representative micrograph showing non-specific expression throughout cNTS neurons in Phox2b-Cre mice (lower panel). (d) Mean (± s.e.m.) number of nicotine infusions after saline or CNO (1 mg/kg, IP) injection in Phox2b-Cre mice expressing DIO-hM3Dq. *P=0.0032, paired t-test; n=6. (e) Mean (± s.e.m.) number of nicotine infusions after saline or CNO injection to Phox2b-Cre mice expressing hM3Dq (n=5). P=0.0660, paired t-test. (f) Mean (± s.e.m.) number of nicotine infusions after saline or CNO injection in Phox2b-Cre mice expressing AAV-EGFP vector (n=5). P=0.2643, paired t-test. (g) Upper panel: DIO vector design. Lower panel: Representative micrograph showing selective expression of AAV-DIO-hM3Dq-mCherry in GLP-1-immunopositive neurons of Gcg-Cre mice. (h) Mean (± s.e.m.) number of nicotine infusions after saline or CNO (3 mg/kg, IP) injection in Gcg-Cre mice expressing FLEX-hM3Dq. *P=0.0011, paired t-test; n=8.

Figure 4. GLP-1 inputs from NTS stimulate IPN neurons

(a) Graphical representation of AAV-EGFP injection into the mouse cNTS and projection of NTS neurons to IPN. (b) Midbrain micrograph following cNTS AAV-EGFP injection. Insert shows cNTS-to-IPN projections (in green; identified by white arrows). (c) Micrograph of IPN showing GLP-1 immunoreactive fibers in IPN (green) and surrounding VTA identified by TH immunoreactivity (red). Insert shows GLP-1+ fiber in IPN (identified by white arrow). (d) Upper panel: DIO-ChR2-GFP vector design. Lower panel: Graphical representation of site of DIO-ChR2-GFP injection into NTS. (e) Representative micrographs showing expression of AAV-DIO-ChR2-GFP in GLP-1-immunopositive neurons. (f) Representative micrographs showing expression of GFP+ fibers in IPN. (g) Example traces of currents evoked by 1 Hz optical stimulation in IPN neurons. (h) Summarized results (mean ± s.e.m.) of changes in relative frequency of excitatory currents in IPN neurons by 1 Hz optical stimulation. (i) Summarized results (mean ± s.e.m.) of changes in relative amplitude of excitatory currents in IPN neurons by 1 Hz optical stimulation. (j) Example traces of currents evoked by 20 Hz optical stimulation in IPN neurons. (k) Summarized results (mean ± s.e.m.) of changes in relative frequency of excitatory currents in IPN neurons by 20 Hz optical stimulation. (l) Summarized results (mean ± s.e.m.) of changes in relative amplitude of excitatory currents in IPN neurons by 20 Hz optical stimulation. Yellow scale bar, 100μm; orange scale bar, 25 μm.

Figure 5. GLP-1 activates IPN neurons by stimulating habenular terminals

(a) Graphical in IPN neurons before and after bath application of Ex-4 (100 nM). (b) Cumulative probability (c) and summarized results (mean ± s.e.m.) showing that the relative amplitude of mEPSCs in IPN neurons is not altered by Ex-4. (d) Cumulative probability (e) and summarized results (mean ± s.e.m.) showing that the relative frequency of mEPSCs in IPN neurons is increased by Ex-4. *P<0.05, paired t-test; n=9 cells from 4 animals. (f) Graphical representation the MHb-IPN circuit (green). Micrographs below: Cholinergic MHb neurons (left) send axonal projections to the IPN (right) as evidenced by fluorescence from ChAT-ChR2-eYFP mice. Nuclear DAPI staining is shown in blue. (g) Sample trace showing that the amplitude of light-evoked EPSCs in IPN neurons from ChAT-ChR2-eYFP mice is increased by Ex-4. (h) Summarized results showing that Ex-4 significantly increases the amplitude of light-evoked EPSC in IPN neurons. ***P<0.001, paired t-test; n=13 cells from 9 cells animals. (i) Representative micrographs showing induction of Fos in IPN following nicotine challenge in GLP-1R KO and wild-type mice. (j) Mean (± s.e.m.) number of Fos-positive neurons per IPN section in GLP-1R knockout (saline, n=6; nicotine, n=7) and wild-type (saline, n=6; nicotine, n=7) mice following nicotine challenge. Two-way ANOVA, Genotype: F_{(1, 22)}=17.69, p=0.0004; Nicotine: F_{(1, 22)}=51.71, p<0.0001; Genotype × Nicotine: F_{(1, 10)}=13.36, **p=0.0014.

Figure 6. GLP-1 transmission in IPN regulates nicotine intake

(a) Graphical representation the MHb-IPN circuit (green) and targeting of AAV-sh-Glp1r-GFP or AAV-GFP viruses into MHb. Micrographs below: GFP+ neurons in MHb neurons and GFP+ fibers in the fasciculus retroflexus (Fr) (left) and GFP+ terminals (right) in virus-treated rats. (b) Mean (± s.e.m.) number of nicotine infusions in AAV-GFP and AAV-sh-Glp1r-GFP rats. *P<0.05, post-hoc test after main effect in two-way ANOVA between intake at the higher nicotine doses (0.09, 0.12 mg kg⁻¹ per infusion); n=6–7 per group). (c) Graphical representation of indwelling cannula directed toward IPN for microinjection (upper panel) and photograph showing representative cannula tract (identified by black arrows) in rat included in study (lower panel). (d) Mean (± s.e.m.) number of nicotine infusions earned after IPN infusion of vehicle (saline) or Ex-4 (0.1μg/0.5 μl) in rats (n=6) pre-injected with vehicle (saline, blue) or cAMPS-Rp (gray). Two-way RM ANOVA, Ex-4: F_{(1, 10)}= 13.1, **p<0.01; cAMPS-Rp: F_{(1, 10)}= 8.56, p<0.05; Ex-4 × cAMPS-Rp: F_{(1, 10)}= 4.3, p=0.06. (e) Mean (± s.e.m.) number of nicotine infusions earned following control microinjection of vehicle (saline) or Ex-4, 2 mm above the IPN (n=6). (f) Mean (± s.e.m.) number of nicotine infusions earned following IPN microinjection of vehicle (saline) or Ex-9 (20 μg; n=9). One-way RM ANOVA: F_{(2, 16)}=4.3, p<0.05. *P<0.05 compared with vehicle; Bonferroni’s test.

Figure 7. GLP-1 in IPN abolishes nicotine reward

(a) Graphical representation of experimental design. (b) Graphical representation of ICSS procedure, in which a rat responds on a wheel manipulandum to receive rewarding intracranial electrical stimulation. The minimally rewarding stimulation intensity that will support reliable responding is termed the reward threshold. (c) Mean (± s.e.m.) percentage change from baseline reward thresholds recorded immediately before (gray; pre-nicotine) and after (blue; post-nicotine) nicotine self-administration in animals receiving intra-IPN infusion of vehicle or Ex-4 (0.1 μg/0.5 μl). Two-way RM ANOVA, Nicotine: F_{(1, 22)}=5.8, p<0.05; Ex-4: F_{(1, 22)}= 11.8, p = 0.01; Nicotine × Ex-4: F_{(1, 22)}=11.3, p<0.01. ***P<0.001 compared with pre-nicotine thresholds; Bonferroni’s test. (d) Mean (± s.e.m.) number of nicotine infusions earned following first ICSS session by rats (n=12) after intra-IPN pretreatment with vehicle (saline) or Ex-4. ***P=0.0002, paired t-test.