Glucose-Dependent Insulinotropic Polypeptide Receptor-Expressing Cells in the Hypothalamus Regulate Food Intake

Abstract

Ambiguity regarding the role of glucose-dependent insulinotropic polypeptide (GIP) in obesity arises from conflicting reports asserting that both GIP receptor (GIPR) agonism and antagonism are effective strategies for inhibiting weight gain. To enable identification and manipulation of Gipr-expressing (Gipr) cells, we created Gipr-Cre knockin mice. As GIPR-agonists have recently been reported to suppress food intake, we aimed to identify central mediators of this effect. Gipr cells were identified in the arcuate, dorsomedial, and paraventricular nuclei of the hypothalamus, as confirmed by RNAscope in mouse and human. Single-cell RNA-seq identified clusters of hypothalamic Gipr cells exhibiting transcriptomic signatures for vascular, glial, and neuronal cells, the latter expressing somatostatin but little pro-opiomelanocortin or agouti-related peptide. Activation of G_q-DREADDs in hypothalamic Gipr cells suppressed food intake in vivo, which was not obviously additive with concomitant GLP1R activation. These data identify hypothalamic GIPR as a target for the regulation of energy balance.

Keywords: food intake; glucose-dependent insulinotropic polypeptide; glucose-dependent insulinotropic polypeptide receptor; hypothalamus.

Graphical abstract

Figure 1

Gipr-Expressing Cells in the Brain (A) Micrograph of GFP staining in brain from heterozygous Gipr^EYFP mice (see also Figure S1). (B) Relative expression of Gipr in whole hypothalamic homogenates in WT mice (n = 3). Data are plotted as 2^ΔCt compared to Actb with the bar representing mean ± SD. (C) Gipr cells were isolated from single-cell digests of hypothalami from two heterozygous Gipr^EYFP mice via FACS, and their transcriptomes were analyzed by scRNA-seq followed by clustering analysis. tSNE visualization of hypothalamic Gipr cells indicates that there are six clusters (top). Cell types were assigned according to expression of a combination of marker genes (bottom) (see also Table S1). (D) t-SNE plots of the expression of selected markers for neurons (Snap25), GABAergic neurons (Slc32a1), glutamatergic neurons (Slc17a6), oligodendrocytes (Mal), mural cells (Abcc9 and Mustn1), VLMCs (Lum), and ependymocytes (Ccdc153). (E) Violin plots representing expression of genes encoding secreted products within the neuronal cluster.

Figure 2

Gipr-Expressing Cells Are Activated by Endocrine Factors (A) Violin plots depicting the expression of GPCRs in cells from the neuronal cluster. (B and C) Ligands for a selection of receptors were tested using calcium imaging in primary cultures of adult hypothalamic cells from heterozygous Gipr^GCaMP3 mice. Dispersed hypothalamic cells were imaged 2–16 h after plating. Cells were perfused with stimuli as indicated. Example traces are shown in (B), and data from all cells tested are represented in (C), with the number of responding cells out of the total number imaged for each condition represented above each bar. Bars represent the mean ± SE.

Figure 3

Activation of Hypothalamic Gipr-Expressing Cells Decreases Food Intake Heterozygous Gipr-Cre mice were injected bilaterally with AAV-DIO-hM3D-mCherry into the hypothalamus to produce Gipr^hypDq mice. CNO (1 mg/kg) or vehicle was injected i.p. following either ad lib feeding or a 10-h daytime fast before dark-phase food intake or following a 2-h fast for light-phase measurements. These paradigms were tested in both chow- (A)–(C) and HFD- (D)–(F) fed mice. Different symbols (squares and circles) indicate mice from different experimental cohorts (see also Figure S3). Dark-phase food intake was compared using a paired t test. Light-phase food intake was compared using a repeated measures 2-way ANOVA with a Sidak’s post-hoc test. ^⁎p < 0.05, ^⁎⁎p < 0.01, ^⁎⁎⁎p < 0.001; n = 5 (A) and (D), 4 (B), 14 (C), 15 (E), and 14 (F).

Figure 4

Partial Cellular Overlap of Gipr and Glp1r Expression, but Limited Effect of GLP1R-Co-activation on Gipr-Expressing Cell-Mediated Acute Anorexia (A–E) Coronal sections of mouse (A–C) and human (D–E) hypothalamus were co-labeled for Gipr or GIPR and Glp1r or GLP1R mRNA using RNAscope. Areas corresponding to the ARC and DMH in mouse and PVH/DMH, lateral hypothalamus (LH), and mediobasal hypothalamus (MBH) in human were assessed for Gipr or GIPR and Glp1r or GLP1R expression (B), (Di), and (Ei). Single- and double-labeled cells were counted and scored (C), (Dii), and (Eii.). Bars represent the mean ± SD (see also Figure S4). (F) Gipr-Cre x Glp1r-Cre and Glp1r-Cre-only mice were injected bilaterally with AAV-DIO-hM3D-mCherry into the hypothalamus to produce Gipr/Glp1r^hypDq and Glp1r^hypDq mice, respectively. CNO (1 mg/kg) or vehicle was injected i.p. following a 10-h daytime fast at the onset of the dark phase before measuring food intake 2 h post-activation (see also Figures S3C and S3D). Food intake was compared using a repeated measures 2-way ANOVA with a Sidak’s post-hoc test. ^⁎⁎p < 0.01, Gipr/Glp1r^hypDq n = 7, Glp1r^hypDq n = 4. (G) Heterozygous Gipr-Cre mice were injected bilaterally with AAV-DIO-hM3D-mCherry into the hypothalamus to produce Gipr^hypDq mice. Following a 10-h daytime fast Exendin-4 (Ex-4) (1.5 nmol/kg) or saline was injected s.c. 1 h prior to the onset of the dark phase. CNO (0.3 mg/kg) or vehicle was injected i.p. at the onset of the dark phase, food was presented, and food intake measurements were taken 2 h post-activation (see also Figures S3E, S4C, and S4D). Bars represent mean ± SD. Food intake was compared using a repeated measures 2-way ANOVA with a Sidak’s post-hoc test. ^⁎p < 0.05, ^⁎⁎p < 0.01, Gipr^hypDq n = 12.