Metabolic and gut microbiome changes following GLP-1 or dual GLP-1/GLP-2 receptor agonist treatment in diet-induced obese mice

Abstract

Enteroendocrine L-cell derived peptide hormones, notably glucagon-like peptide-1 (GLP-1) and glucagon-like peptide-2 (GLP-2), have become important targets in the treatment of type 2 diabetes, obesity and intestinal diseases. As gut microbial imbalances and maladaptive host responses have been implicated in the pathology of obesity and diabetes, this study aimed to determine the effects of pharmacologically stimulated GLP-1 and GLP-2 receptor function on the gut microbiome composition in diet-induced obese (DIO) mice. DIO mice received treatment with a selective GLP-1 receptor agonist (liraglutide, 0.2 mg/kg, BID) or dual GLP-1/GLP-2 receptor agonist (GUB09-145, 0.04 mg/kg, BID) for 4 weeks. Both compounds suppressed caloric intake, promoted a marked weight loss, improved glucose tolerance and reduced plasma cholesterol levels. 16S rDNA sequencing and deep-sequencing shotgun metagenomics was applied for comprehensive within-subject profiling of changes in gut microbiome signatures. Compared to baseline, DIO mice assumed phylogenetically similar gut bacterial compositional changes following liraglutide and GUB09-145 treatment, characterized by discrete shifts in low-abundant species and related bacterial metabolic pathways. The microbiome alterations may potentially associate to the converging biological actions of GLP-1 and GLP-2 receptor signaling on caloric intake, glucose metabolism and lipid handling.

Conflict of interest statement

M.S.M., P.W., K.F., K.R., M.M. and H.H.H. are employed by Gubra; J.B.H., A.P. and H.B.N. are employed by Clinical Microbiomics; N.V. and J.J. are owners of Gubra.

Figures

Figure 1

Liraglutide and GUB09-145 improve metabolic parameters in DIO mice. (A) Absolute body weight (g), (B) Body weight gain (%) relative to treatment start; (C) Daily food intake (g); (D) Cumulative energy intake (kcal/day); (E) Oral glucose tolerance test (OGTT) on treatment day 27, (F) Glucose area-under the-curve (glucose AUC0-240 min); (G) Fasting blood glucose concentrations (mmol/L) on treatment day 14; (H) Fasting plasma insulin levels (pg/ml) on treatment day 28; (I) Plasma total triglycerides (TG, mmol/L) on treatment day 28; (J) Plasma total cholesterol (TC, mmol/L) on treatment day 28. *p < 0.05, **p < 0.01, ***p < 0.001 compared to DIO vehicle controls.

Figure 2

16S rDNA sequencing. Principal Coordinate Analysis (PCoA) based on Bray-Curtis dissimilarity between all samples (panels A,B). PCoA plots show pre- and post-treatment gut microbiome signatures in all individual mice, group means are indicated by a large point. Microbial diversity analysis at genus level, including sample (OTU) richness (C) and Shannon alpha diversity (D). *p < 0.05, **p < 0.01, ***p < 0.001 vs. corresponding sampling time in DIO vehicle mice; #p < 0.05, ##p < 0.01 vs. corresponding baseline.

Figure 3

Taxonomic summary at phylum and family level. Mean values per group are illustrated of the nine detected phyla (panel A) and the 15 most abundant families (panel B) across all samples. The phyla and families are sorted with the highest abundance across all samples in the top. For panel B, note that bar height for liraglutide and GUB09-145 (endpoint data) is slightly lower compared to baseline, reflecting that terminal samples showed a relatively higher proportion of reads not mapping to the top-15 most abundant families.

Figure 4

Shotgun metagenomics. Principal Coordinate Analysis (PCoA) based on Bray-Curtis dissimilarity between all samples calculated using species relative abundances (panels A,B). PCoA plots demonstrate pre- and post-treatment gut microbiome signatures in all individual mice, group means are indicated by a large point. Microbial diversity analysis at species level, including gene richness (C) and Shannon alpha diversity (D). ***p < 0.001 vs. corresponding sampling time in DIO vehicle mice.

Figure 5

Correlation of metagenomic species abundance to changes in metabolic parameters. Overview of metagenomic species significantly changing in abundance over the course of liraglutide and GUB09-145 treatment. Hierarchical clustering analysis yielded four individual species categories: (Cluster 1) species with reduced abundance after liraglutide or GUB09-145 treatment; (Cluster 2) species with increased abundance primarily after liraglutide treatment; (Cluster 3) species with abundance change specific to one drug treatment group only; (Cluster 4) species with increased abundance after both liraglutide and GUB09-145 treatment, and no differences between the two vehicle groups. Left heatmap: Spearman correlations between change in individual species abundance and various metabolic parameters affected by treatment, including plasma total cholesterol (TC), fasting glucose level on treatment day 14 (Fasting glucose), fasting terminal insulin levels (fasting insulin), glucose area-under the curve in an oral glucose tolerance test on treatment day 27 (AUC glucose OGTT), terminal plasma total triglycerides (TG), endpoint body weight loss relative to baseline (Body weight loss), and total energy intake during the treatment period (Total caloric intake). Asterisks denote significant correlation (FDR-corrected, p < 0.05) between individual species and relevant metabolic parameter. Right heatmap: Fold change (log2 transformed) in species abundance in individual mice as compared to baseline.

Figure 6

Bacterial KEGG metabolic pathways affected by treatment with liraglutide and GUB09-145. Heatmap depicting bacterial KEGG modules with changes over the course of liraglutide and GUB09-145 treatment. Left heatmap: Spearman correlations between change in individual species abundance and various metabolic parameters affected by treatment, including plasma total cholesterol (TC), fasting glucose level on treatment day 14 (Fasting glucose), fasting terminal insulin levels (fasting insulin), glucose area-under the curve in an oral glucose tolerance test on treatment day 27 (AUC glucose OGTT), terminal plasma total triglycerides (TG), endpoint body weight loss relative to baseline (Body weight loss), and total caloric intake during the treatment period (Total caloric intake). Asterisks denote significant correlation (FDR-corrected, p < 0.05) between species and relevant metabolic parameter. Right heatmap: Fold change (log2 transformed) in species abundance in individual mice as compared to baseline.