Impact of metformin and Dysosmobacter welbionis on diet-induced obesity and diabetes: from clinical observation to preclinical intervention

Aims/hypothesis We aimed to investigate the association between the abundance of Dysosmobacter welbionis, a commensal gut bacterium, and metabolic health in human participants with obesity and diabetes, and the influence of metformin treatment and prebiotic intervention. Methods Metabolic variables were assessed and faecal samples were collected from 106 participants in a randomised controlled intervention with a prebiotic stratified by metformin treatment (Food4Gut trial). The abundance of D. welbionis was measured by quantitative PCR and correlated with metabolic markers. The in vitro effect of metformin on D. welbionis growth was evaluated and an in vivo study was performed in mice to investigate the effects of metformin and D. welbionis J115T supplementation, either alone or in combination, on metabolic variables. Results D. welbionis abundance was unaffected by prebiotic treatment but was significantly higher in metformin-treated participants. Responders to prebiotic treatment had higher baseline D. welbionis levels than non-responders. D. welbionis was negatively correlated with aspartate aminotransferase (AST) and alanine aminotransferase (ALT) levels and fasting blood glucose levels in humans with obesity and type 2 diabetes. In vitro, metformin had no direct effect on D. welbionis growth. In mice, D. welbionis J115T treatment reduced body weight gain and liver weight, and improved glucose tolerance to a better level than metformin, but did not have synergistic effects with metformin. Conclusions/interpretation D. welbionis abundance is influenced by metformin treatment and associated with prebiotic response, liver health and glucose metabolism in humans with obesity and diabetes. This study suggests that D. welbionis may play a role in metabolic health and warrants further investigation. Clinical trial NCT03852069 Graphical Abstract Supplementary Information The online version contains peer-reviewed but unedited supplementary material available at 10.1007/s00125-023-06032-0.

Participants were included for a period of three months and randomized to consume either 16g/d of native inulin (extracted from chicory root, Cosucra, Belgium) or 16 g/d of maltodextrin (Cargill, Belgium), provided in identical packaging.Empty and unused packets were returned to measure compliance.To promote adaptation to the fiber, patients were asked to ingest half the dose during the first week.Participants received a cookbook with recipes based on vegetables either rich or poor in fructans and were advised to consume at least one meal proposed in the recipe book per day.Stools samples were collected at baseline and at the end of the 3-month of intervention and stored at room temperature with a DNA stabilizer (Stratec biomolecular, Berlin, Germany) for maximum three days, then transferred to -80°C.Genomic DNA was extracted from faeces using a PSP® spin stool DNA kit (Stratec biomolecular, Berlin, Germany).Blood samples were collected using BD TM P800 Collection and Preservation System, which contains DPP-IV and other protease inhibitors.Plasma concentrations of Fibroblast growth factor 21 (FGF21), active ghrelin, total glucagon-like peptide 1 (GLP-1), glucagon, leptin, pancreatic polypeptide (PP), total peptide YY (PYY) were determined using the Meso Scale Discovery (MSD) U-PLEX assay (Rockville, MD, USA) following the manufacturer's instructions.Analyses were performed using a QuickPlex SQ 120 instrument (MSD) and DISCOVERY WORKBENCH® 4.0 software (MSD, Rockville, MD, USA).qPCR quantification of Dysosmobacter welbionis and total bacteria in human samples Absolute quantification of the total bacterial load was performed by quantitative polymerase chain reaction (qPCR) using the primers Bacteria Universal P338F (ACTCCTACGGGAGGCAGCAG) and P518R (ATTACCGCGGCTGCTGG) as previously described in [4].Real-time PCR was performed with a QuantStudio3 (Applied Biosystems, The Netherlands) using SYBR Green (GoTaq® qPCR mix, Promega, USA) for detection and using the QuantStudio Software (Version 1.4.3,Applied Biosystems, The Netherlands).Absolute quantification was achieved through the inclusion of a standard curve (performed in duplicate) on each plate generated by diluting DNA from pure culture of L. acidophilus NCFM (five-fold serial dilution).Cell counts were determined by plating and expressed as "colony-forming unit" (CFU) before DNA isolation.Culture and preparation of Dysosmobacter welbionis for mouse experiments D. welbionis J115 T was cultured anaerobically in a modified YCFA medium supplemented with 10 g/L inositol.Cultures were centrifuged at 5000g during 15 min and the supernatant was removed.Cells were then resuspended in anaerobic PBScarbonate buffer supplemented with 15 % (vol/vol) trehalose, then immediately frozen in anaerobic vials and stored at -80°C.The number of total and cultivable bacteria administered to the mice was calculated by plating the bacterial culture before preparation and the bacterial suspension after preparation for administration.

Mouse model
Sets of 7-week-old C57BL/6J male mice (Janvier Laboratories, Le Genest-Saint-Isle, France) were housed in pairs in SOPF (specific opportunistic and pathogen free) conditions, in a controlled environment (room temperature of 22 ± 2°C, 12h daylight cycle) with free access to sterile (irradiated) food and sterile (autoclaved) water.Upon arrival, mice were randomly separated at 2 animal/cage and underwent a 1-week acclimatization period, during which they were fed a control diet (CT (AIN93Mi, Research Diet, New Brunswick, NJ, USA).During the experiments, food and water intake were recorded once a week.Body composition was assessed by using a 7.5 MHz time domain-nuclear magnetic resonance machine (TD-NMR) (LF50 minispec, Bruker, Rheinstetten, Germany).
A set of 50 mice was divided in 5 groups of 10 mice.The mice were fed either control diet (CT, AIN93Mi) or high-fat diet (HFD, 60% fat and 20% carbohydrates (kcal/100g), D12492, Research diet).The HFD+J115 group was given Dysosmobacter welbionis J115 T by oral gavage at the dose 1.10 9 cfu/0.2ml per day and per mice.The HFD+metformin group was orally administered 150 mg/kg/day the first 2 weeks and 100 mg/kg/day for the rest of the study.In the group HFD+J115+metformin, mice were treated with joint administration of Dysosmobacter welbionis J115 T and metformin by daily oral gavage.In the control groups, control diet (CT) and HFD mice were treated with an oral gavage of an equivalent volume of PBS-carbonate buffer supplemented with 15% (weight/vol) trehalose.The treatment continued for 10 weeks.Mice were killed after a 6-hour fasting period.

Oral glucose tolerance test (OGTT)
One week before the end of the experiment, mice were fasted for 6 hours before being given an oral gavage glucose load (2 g glucose per kg body weight).Blood glucose was measured 30 minutes before (timepoint -30), just prior to the oral glucose load (timepoint 0) and then after 15, 30, 60, 90 and 120 min.Blood glucose was determined with a glucometer (Accu Chek Performa, Roche, Basel, Switzerland) on blood samples collected from the tip of the tail vein.Plasma insulin concentration was determined by ELISA (Mercodia, Uppsala, Sweden) according to the manufacturer's instructions.

Tissue sampling
The animals were anesthetized with isoflurane (Forene ® , Abbott, Queenborough, Kent, England) and blood was collected from the portal and cava veins.Then mice were killed by decapitation.Tissue samples (liver, brown adipose tissue, subcutaneous adipose tissue, mesenteric adipose tissue) were dissected, immersed in liquid nitrogen, and stored at -80 °C for further analysis.A part of the adipose tissues was fixed in 4 % paraformaldehyde in PBS for histological analysis.

Insulin resistance index
The insulin resistance index was determined by multiplying the area under the curve (from −30 to 15 min) of blood glucose and plasma insulin obtained during the OGTT.

Histological analyses
Brown and white adipose tissues were fixed in 4 % paraformaldehyde for 24 h at room temperature.Samples were then dehydrated by immersion in ethanol 100% for 24 h and processed for paraffin embedding.Paraffin sections of 5 μm were stained with haematoxylin and eosin.Whole tissue sections were digitalized using a Panoramic ScanII slide scanner (3DHistech Ltd, Budapest, Hungary) with a 20x Plan-Apochromat objective and visualized with the Cytomine web platform.The white area in brown adipose tissue corresponding to (empty) lipid droplets were quantified based on 5 fields per sample using Fiji software.The adipocyte size and frequency in SAT and VAT was assessed using Adiposoft [5] on Fiji software.5 pictures per adipose tissue were taken.

Gene expression analysis by real-time qPCR
Total RNA was prepared from tissues using TriPure reagent (Roche).Quantification and integrity analysis of total RNA was performed by running 1 μl of each sample on an Agilent 2100 Bioanalyzer (Agilent RNA 6000 Nano Kit, Agilent).
For qPCR analysis, cDNA was prepared by reverse transcription of 1 μg total RNA using a Reverse Transcription System kit (Promega, Leiden, The Netherlands Animal ethics All mouse experiments were approved by the Ethical Committee for Animal Care of the Health Sector of the Université Catholique de Louvain (UCLouvain) headed by Prof. J-P Dehoux, under number 2022/UCL/MD/09, and were performed in accordance with the guidelines of the Local Ethics Committee and in accordance with the Belgian Law of 29 May 2013, regarding the protection of laboratory animals (agreement number LA1230314)

Table 1 . Primers for qPCR
). Realtime PCRs were performed with the StepOnePlus real-time PCR system and software (Applied Biosystems, Den Ijssel, The Netherlands) using GoTaq® qPCR mix, (Promega, USA) for detection, according to the manufacturer's instructions.RPL19 was chosen as the reference gene.All samples were run in duplicate in a single 96well reaction plate, and data were analyzed according to the 2-ΔΔCt method.The identity and purity of the amplified product were checked through analysis of the melting curve at the end of amplification.Primer sequences for the mouse genes are shown in supplemental ESM Table1.
considered statistically significant.The presence of outliers was assessed using ROUT's outlier test on GraphPad Prism.Figure1d-e and S2: We used Rstudio program to perform the correlation matrices using tidyverse, dplyr, ggplot2, corrr, ESM