Behavioural and neurochemical mechanisms underpinning the feeding-suppressive effect of GLP-1/CCK combinatorial therapy

Abstract

Objectives: Combinatorial therapies are under intense investigation to develop more efficient anti-obesity drugs; however, little is known about how they act in the brain to produce enhanced anorexia and weight loss. The goal of this study was to identify the brain sites and neuronal populations engaged during the co-administration of GLP-1R and CCK1R agonists, an efficient combination therapy in obese rodents.

Methods: We measured acute and long-term feeding and body weight responses and neuronal activation patterns throughout the neuraxis and in specific neuronal subsets in response to GLP-1R and CCK1R agonists administered alone or in combination in lean and high-fat diet fed mice. We used PhosphoTRAP to obtain unbiased molecular markers for neuronal populations selectively activated by the combination of the two agonists.

Results: The initial anorectic response to GLP-1R and CCK1R co-agonism was mediated by a reduction in meal size, but over a few hours, a reduction in meal number accounted for the sustained feeding suppressive effects. The nucleus of the solitary tract (NTS) is one of the few brain sites where GLP-1R and CCK1R signalling interact to produce enhanced neuronal activation. None of the previously categorised NTS neuronal subpopulations relevant to feeding behaviour were implicated in this increased activation. However, we identified NTS/AP Calcrl⁺ neurons as treatment targets.

Conclusions: Collectively, these studies indicated that circuit-level integration of GLP-1R and CCK1R co-agonism in discrete brain nuclei including the NTS produces enhanced rapid and sustained appetite suppression and weight loss.

Keywords: CCK; Combinatorial therapy; GLP-1; Hunger; Nucleus of the solitary tract; Obesity; Satiety.

Figure 1

A subthreshold dose of GLP-1R agonist AC3174 acutely increased the anorectic response to CCKR1 agonism. Two h food intake in response to an i.p. injection of saline, CCK-8, exenatide, GLP-1R agonist AC3174, CCK1R agonist AC170222, CCK2R agonist AC170236, or an equimolar combination of these drugs in the lean chow-fed mice (A–D) or the mice fed a HF diet for 3 weeks (E). Data (means ± sem) were analysed using 2-way ANOVA for repeated measures with 2 factors (treatment and time) and Tukey's multiple comparison tests. The mice were 9 weeks old at the beginning of the studies (±26 g).

Figure 2

Chronic GLP-1R and CCK1R co-agonism produced sustained hypophagia and weight loss. The mice were treated with i.p. administration of saline, GLP-1R agonist AC3174 (3 μg/kg), CCK1R agonist AC170222 (30 μg/kg), or AC3174 + AC170222 twice daily for 5 days. Arrows indicate injections and stars indicate times at which body weight was measured (A). Night-time food intake (B), weight loss (C), and hourly night-time food intake during night 1 (D). Data (means ± sem) were analysed using 2-way ANOVA for repeated measures with 2 factors (treatment and time) and Tukey's multiple comparison tests. The mice were 9 weeks old at the beginning of the studies (±26 g).

Figure 3

Neuronal activation in response to acute GLP-1R and CCK1R co-agonism in hypothalamic and caudomedial hindbrain sites. C-fos immunolabelling in the AP (A), NTS (B), ARH (C), VMH (D), PVH (E), and DMH (F) 80 min after an i.p. administration of saline, GLP-1R agonist AC3174, CCK1R agonist AC170222, or AC3174 + AC170222. Data are means ± sem. ∗p < 0.05, ∗∗p < 0.01, and ∗∗∗∗p < 0.0001 vs saline. The mice were 9 weeks old at the beginning of the studies (±26 g). A minimum of 4 sections per mouse per brain region were analysed.

Figure 4

GLP-1R and CCK1R co-agonism did not integrate within CCK1R or GLP-1R neurons. Representative images showing Cckar (A and D) and Glp-1r (B and C) distribution in the AP, NTS, and ARH. Representative images and quantification of neuronal activation in AP (E), ARH (F), or GLP-1R neurons in response to an i.p. administration of saline, GLP-1R agonist AC3174 (1 μg/kg), CCK1R agonist AC170222 (2.5 μg/kg), or AC3174 + AC170222 (1 μg/kg + 2.5 μg/kg). Data are means ± sem. ∗p < 0.05 vs saline.

Figure 5

Neurochemical characterisation of the neuronal subpopulations activated by the combination of GLP-1R and CCK1R agonists. Representative images and quantification of neuronal activation in NTS POMC (A), AP POMC (B), ARH POMC (C), NTS TH (D), AP TH (E), and ARH NPY (F) neurons in response to an i.p. administration of saline, GLP-1R agonist AC3174 (1 μg/kg), CCK1R agonist AC170222 (2.5 μg/kg), or AC3174 + AC170222 (1 μg/kg + 2.5 μg/kg). Data are means ± sem. ∗p < 0.05, ∗∗p < 0.01, and ∗∗∗p < 0.001 vs saline. rNTS: rostral NTS; cmNTS: caudomedial NTS.

Figure 6

PhosphoTRAP assay identified the molecular profile and neurons activated or inhibited by the combinatorial treatment in the DVC. Enrichment ratios of transcript significantly (p < 0.01) enriched in neurons activated (A) and inhibited (B) by co-treatment with GLP-1R agonist AC3174 and CCK1R agonist AC170222, but not in neurons activated or inhibited by the monotherapies (AC3174 vs saline: p > 0.1 and AC170222 vs saline: p > 0.1). Differential enrichment ratios (C) of transcripts enriched in neurons more activated by the combinatorial therapy than monotherapies (AC3174 vs combo: p < 0.05 and AC170222 vs combo: p < 0.05). Differential enrichment ratios of Calcrl (D). Expression of Calcrl in the caudal medulla (E). Co-localisation of Cckar (green), Glp1r (white), and Calcrl (red) in the DVC (F). Co-localisation of c-fos (green) and Calcrl (red) in response to i.p. administration of the vehicle (left) or combinatorial treatment in the chow-fed mice (right) (G). Expression of c-fos (green) (H) and co-localisation of c-fos (green) and Calcrl (red) (I) in response to i.p. administration of the vehicle, AC3174, AC170222, or combinatorial treatment in the high-fat-fed mice. Feeding response to saline or AC3174 (1 μg/kg) + AC170222 (10 μg/kg) in the mice pre-treated with vehicle or AC187 (100 μg/kg) (J). Data are means ± sem. Means sharing a letter are not statistically different.

Figure 7

PhosphoTRAP assay identified the molecular profile and neurons activated or inhibited by the combinatorial treatment in the MBH. Enrichment ratios of transcripts significantly (p < 0.01) enriched in neurons activated (A) and inhibited (B) by co-treatment with the GLP-1R agonist AC3174 and CCK1R agonist AC170222, but not in neurons activated or inhibited by the monotherapies (AC3174 vs saline: p > 0.1 and AC170222 vs saline: p > 0.1). Differential enrichment ratios (C) of transcripts with higher enrichment in response to the combinatorial therapy than the monotherapies (AC3174 vs combo: p < 0.05 and AC170222 vs combo: p < 0.05). RNAscope ISH against c-fos (D and F) and c-fos + Adcyap1r1 (E and F) in the ARH of the mice treated with the combinatorial treatment vehicle. Data are means ± sem. ∗p < 0.05.