Loss of GIPR in LEPR cells impairs glucose control by GIP and GIP:GLP-1 co-agonism without affecting body weight and food intake in mice

Abstract

Objective: The glucose-dependent insulinotropic polypeptide (GIP) decreases body weight via central GIP receptor (GIPR) signaling, but the underlying mechanisms remain largely unknown. Here, we assessed whether GIP regulates body weight and glucose control via GIPR signaling in cells that express the leptin receptor (Lepr).

Methods: Hypothalamic, hindbrain, and pancreatic co-expression of Gipr and Lepr was assessed using single cell RNAseq analysis. Mice with deletion of Gipr in Lepr cells were generated and metabolically characterized for alterations in diet-induced obesity (DIO), glucose control and leptin sensitivity. Long-acting single- and dual-agonists at GIPR and GLP-1R were further used to assess drug effects on energy and glucose metabolism in DIO wildtype (WT) and Lepr-Gipr knock-out (KO) mice.

Results: Gipr and Lepr show strong co-expression in the pancreas, but not in the hypothalamus and hindbrain. DIO Lepr-Gipr KO mice are indistinguishable from WT controls related to body weight, food intake and diet-induced leptin resistance. Acyl-GIP and the GIPR:GLP-1R co-agonist MAR709 remain fully efficacious to decrease body weight and food intake in DIO Lepr-Gipr KO mice. Consistent with the demonstration that Gipr and Lepr highly co-localize in the endocrine pancreas, including the β-cells, we find the superior glycemic effect of GIPR:GLP-1R co-agonism over single GLP-1R agonism to vanish in Lepr-Gipr KO mice.

Conclusions: GIPR signaling in cells/neurons that express the leptin receptor is not implicated in the control of body weight or food intake, but is of crucial importance for the superior glycemic effects of GIPR:GLP-1R co-agonism relative to single GLP-1R agonism.

Keywords: GIP; GIPR:GLP-1R co-agonism; GLP-1; Obesity; Type 2 diabetes.

Figure 1

Single nuclei (sn)RNA-seq analysis of Lepr positive cells in the murine hypothalamus. Schematic for the selective isolation of hypothalamic Lepr positive nuclei from Lepr-Cre Sun1-GFP-Myc reporter mice using fluorescence-activated cell sorting (FACS) (A). Feature plots of snRNAseq for cells expressing Lepr, Gipr, and Lepr/Gipr(B). The overlay panel reveals low abundance for Gipr that appears restricted to the clusters St18, Gpc6, Nxph1 and Ntm. Gipr is largely absent from the Pomc and Agrp clusters. Feature plots for the glutamatergic (C) and GABAergic (D) neuronal markers Slc17a6 and Slc32a1, respectively. Pie charts of total and relative numbers of Gipr positive nuclei (E–H) in all clusters combined (E), Agrp and Pomc clusters (F), glutamatergic (Slc17a6) vs. GABAergic (Slc32a1) neurons (G) and clusters with low but detectable Gipr abundance, defined via their distinct marker transcripts Suppression of tumorigenicity 18 (St18), Glypican 6 (Gpc6), Neurexophilin 1 (Nxph1) and Neurotrimin (Ntm) (H).

Figure 2

Single cell (sc)RNAseq analysis of Gipr and Lepr in the hindbrain. Expression of Lepr and Gipr in Chat, GABAergic and Glutamatergic neuron clusters in the DVC (A), UMAP visualization of those clusters (B) and the percentage of cells expressing both receptors in DVC cell types (C). UMAP visualization of the relative expression of Lepr(D) and Gipr(E) in DVC cells.

Figure 3

Single cell (sc)RNAseq analysis of Gipr and Lepr in the embryonic murine pancreas. Expression of Lepr(A) and Gipr(B) and percentage of cells expressing Gipr, Lepr, or both receptors within identified clusters (C) in the pancreas of mice at the embryonic ages E12.5 - E18.5 integrated into one data set [17].

Figure 4

Metabolic phenotype of HFD-fed male Lepr-Gipr KO mice. Representative image (A) and quantification (B) of cFos in the hypothalamic arcuate nucleus (ARC) of 19-wk old HFD-fed POMC-GFP mice treated with a single s.c. bolus of acyl-GIP (30 nmol/kg) (n = 5 each group, scale bar 100 μm). Expression of Gipr in the ARC (n = 4–6 each group) (C) and pancreas of 24–32-wk old mice (n = 7 each group) (D). Body weight (E) and body composition of 40-wks old mice (n = 10 each group) (F, G). Cumulative food intake (H), energy expenditure (I), locomoter activity (J), fatty acid oxidation (K), and respiratory exchange ratio (L) of 27-wk old mice (n = 10 each group). Intraperetoneal (i.p.) glucose tolerance of 40-wk old mice (n = 10 each group) (M) and i.p. insulin tolerance in 47-wk old mice (n = 10 each group) (N). Fasting levels of blood glucose in 40-wk old mice (n = 10 each group) (O), and fasting insulin in 47-wk old mice (n = 9 each group) (P). HbA1c in 42-wk old mice (n = 8 each group) (Q) in 40-wk old mice (n = 10 each group), and oral glucose-stimulated insulin secretion in 31-wks old mice (n = 8 each group) (R). Glucose-stimulated insulin secretion in pancreatic islets isolated from 16 to 20-wks old mice and treated with acyl-GLP-1 (50 nM) or acyl-GIP (50 nM) for 45 min under conditions of high (20 mM) glucose (n = 15–20 mice each group). Plasma levels of triglycerides (S), cholesterol (T), and NEFA (U) in 40-wk old mice (n = 10 each group). Data represent means ± SEM. Asterisks indicate ∗ p < 0.05, ∗∗p < 0.01 and ∗∗∗p < 0.001. Longitudinal data (E, H, M, N and R) were analyzed using 2-way ANOVA with time and genotype as co-variables and Bonferroni post-hoc analysis for individual time-points. Bar graphs (B-D, F, G, J-L, O-Q and S–U) were analyzed using 2-tailed, 2-sided ttest. Data in (J) were analyzed using ANCOVA with body weight as co-variate as previously suggested [24,25]. Panel A is a representative example of N = 5 biological replicates. Data points in panels B-D, F,G, I-L, O-Q and S–U represent independent biological replicates.

Figure 5

Metabolic phenotype of chow-fed male Lepr-Gipr KO mice. Body weight (A) and food intake (B) of male Lepr Cre Gipr KO mice and wildtype controls (n = 7–8 each group). Body composition (C, D) at the age of 45-wks (n = 7–8 each group). Intraperetoneal glucose tolerance at the age of 46-wks (n = 7–8 each group) (E) and i.p. insulin tolerance at the age of 49-wks (n = 7–8 each group) (F). Fasting levels of blood glucose in 46-wk old mice (n = 7–8 each group) (G) and of insulin in 48-wk old mice (n = 7 each group) (H). HbA1c (I) and fasting level of total GLP-1 (J) and glucagon (K) in 43-wk old mice (n = 7–8 each group). Plasma levels of triglycerides (L) and NEFA (M) in 47-wk old mice (n = 7–8 each group). Hypothalamic expression of Pomc, Cart, Agrp and Npy in 51-wk old mice (n = 6–8 each group) (M–Q). Data represent means ± SEM. Asterisks indicate ∗ p < 0.05. Longitudinal data (A, B, E, F) were analyzed using 2-way ANOVA with time and genotype as co-variables and Bonferroni post-hoc analysis for individual time-points. Bar graphs (C, D, G-Q) were analyzed using 2-tailed, 2-sided ttest. Cumulative food intake (B) was assessed per cage in single or double-housed mice (n = 7–8 each group). Data points in panels C,D,G-Q represent independent biological replicates.

Figure 6

Metabolic effects of GIP and GIPR:GLP-1R co-agonism in male DIO Lepr-Gipr KO mice. Drug effects in wildtype (A–H) and Lepr Cre Gipr KO mice (I–P). Body weight (A) and food intake (B) of DIO wildtype mice treated daily with acyl-GIP (100 nmol/kg) or 10 nmol/kg of either acyl-GLP-1 or MAR709 (n = 7–8 each group). Change in body composition (n = 6–8 each group) (C, D), i.p. glucose tolerance (E, F), as well as fasting levels of blood glucose (G) and insulin (H) after 23 days of treatment (E-G: n = 7–8 each group; H: 6-8 each group). Body weight (I) and food intake (J) of DIO Lepr-Cre Gipr KO mice treated daily with acyl-GIP (100 nmol/kg) or 10 nmol/kg of either acyl-GLP-1 or MAR709 (n = 6–8 each group). Change in body composition (K, L), i.p. glucose tolerance (M, N), as well as fasting levels of blood glucose (O) and plasma insulin (P) after 23 days of treatment (I–O: n = 6–8 mice each group; P: n = 5–8 mice each group). Food intake (B, I) was assessed per cage in double-, or single-house mice (n = 6–8 mice each group). Data represent means ± SEM. Asterisks indicate ∗ p < 0.05, ∗∗p < 0.01 and ∗∗∗p < 0.001. Longitudinal data (A, E, I, M) were analyzed using 2-way ANOVA with time and treatment as co-variables and using Bonferroni post-hoc analysis for individual time-points. Bar graphs (B-D, F–H, j-l, N–P) were analyzed using 2-tailed, 2-sided ttest for comparison of acyl-GIP vs. Vhcl, and using 1-way ANOVA for comparison of Vhc, acyl-GLP-1 and MAR709. Data points in panels B-D, F–H, J-L, N–P represent independent biological replicates.

Figure S1

Co-expression of Lepr and Gipr based on the HypoMap Repository.

Figure S2

Amino acid sequence of the used peptides. Amino acid sequence and fatty acid acylation of acyl-GIP (IUB0271), acyl-GLP-1 (IUB1746) and the GIPR:GLP-1R co-agonist MAR709.

Figure S3

Plasma and qPCR analysis in male DIO Lepr-Cre Gipr.KO mice. Intraperitoneal Insulin tolerance (A), as well as fasted plasma levels of total GIP (B), total GLP-1 (C), and glucagon (D) in 47-wk old mice (n=8-10 each group). Hypothalamic expression of Pomc(E), Cart(F), Agrp(G) and Npy(H) in 47-wk old mice (n=9-10 each group). Body weight (I), cumulative food intake (J), as well as fat and lean tissue mass (K and L) of 31-wk old DIO Lepr Cre Gipr KO and WT mice treated daily with human recombinant leptin (1 mg/kg/day) (n=8 each group). Data represent means ± SEM. Asterisks indicate ∗ p<0.05, ∗∗ p<0.01 and ∗∗∗ p<0.001. Longitudinal data in panels A, I, and J were analyzed using 2-way ANOVA and Bonferroni post-hoc analysis for individual time points. Bar graphs in panels B-H and K, L were analyzed using 2-tailed, 2-sided ttest. Data points in panels B-H, K, L represent independent biological replicates.