The Interaction Between Protein Intake, Gut Microbiota and Type 2 Diabetes in Subjects With Different Ethnic Backgrounds

Context and justification:

There is growing evidence that the gut microbiota is a key element in the pathophysiology of cardio-metabolic diseases (CMD) such as Type 2 Diabetes (T2D). One hypothesis is that gut-derived metabolites (from diet) have an important role in the host metabolism. Preliminary results show that imidazole propionate (ImP), a degradation product of the essential amino acid histidine, is produced by the gut microbiota of T2D patients, but not healthy subjects. The gut microbiota itself is strongly influenced by diet and ethnicity. However, most dietary intervention studies have focused on the role of fiber intake and the effect of dietary protein on the gut microbiota composition and metabolite production is not well known. Moreover, it has been shown that the response to a dietary intervention may depend on the baseline gut microbiome richness.

Main hypothesis: Depending on the baseline gut microbiome composition, a diminution in protein intake could decrease the microbial production of metabolites such as ImP and improve the metabolism of the host. We also hypothesize that the effects of such an intervention could depend the ethnic background.

Study population:

Individuals with type 2 diabetes (T2D), of Caucasian or Caribbean origin, 120 patients will be included in total

Intervention:

Assignment after randomization to one of the following 2 diets:

• HP diet: high protein (High Protein, HP) diet with 30% protein, 40% carbohydrate and 30% fat (as% of total energy intake)
• LP diet: low protein diet (Low Protein, LP) with 10% protein, 55% carbohydrate and 35% fat (as% of total energy intake) Food boxes adapted to each diet (HP or LP pre-cooked meals, HP or LP breads and snacks) will be provided to the participants throughout the study reaching 40-50% of their prescribed daily energy intake for 6 days per week.

Subjects are required to keep food diaries three days a week and will also have weekly contact with a dietician.

Visits:

• Inclusion visit V0 (maximum 1 month before V1): Participants will first be recruited from the diabetic population of the French cohort of the European project METACARDIS. Individuals eligible for the study are screened for inclusion.

Baseline phenotyping is performed (metabolic, inflammatory blood markers, stool and oral microbiota sampling, body composition by DXA, questionnaires)

• Randomization visit V1 (T0 - start of the intervention): Randomization into 2 parallel groups (High Protein or Low protein diet) will be stratified based on metagenomics richness (obtained from Metacardis results), age (< or ≥ 60), gender, ethnic background (Caribbean or Caucasian).

Meal tolerance test, anthropometric measures, resting energy expenditure measure, one week CGMS, 24h urinary urea measure are performed.

• Follow-up visit V2 (T42 +/- 7 days): Mid protocol visit with anthropometric measures, one week CGMS, 24h urinary urea measures, stool sampling.
• End of study visit V3 (T84 +/- 7 days): Phenotyping is performed (metabolic, inflammatory blood markers, stool and oral microbiota sampling, body composition by DXA, questionnaires, meal tolerance test, anthropometric measures, resting energy expenditure measure, one week CGMS, 24h urinary urea measure)

Statistical analysis:

There are no multiple hypotheses since our study has only one primary objective (AUC delta of the glycemic excursion after a mixed meal tolerance test (MMT) between the beginning of study and 3 months post intervention). Thus, the problem of the type 1 error will not arise.

The primary endpoint will be analyzed to compare changes in AUC for glycemic excursion versus diet (rich vs. low protein), based on initial metagenomic richness (high vs. low) and ethnicity (Caucasian vs. Caribbean). AUC changes after dietary intervention between the different groups will be tested using linear regression models for repeated measurements with adjustment for initial levels. The effect of diet composition within the groups will be tested using Bonferroni's post-hoc covariance analysis (ANCOVA) analyzes. For secondary endpoints, the same approaches will be used for analysis of postprandial metabolites (AUC, AUC, post-MMT variation). Differences in relative abundance of bacterial species and functional modules (generated by metagenomic sequencing) and quality of life questionnaires will also be analyzed by subgroups using uni / multivariate analyzes. Correlations will be sought between changes in bio-clinical variables and changes in measurements of different metabolites.

Funding:

• European Program (Join Program Initiative, JPI HDHL / ERA-NET cofund HDHL-INTIMIC, -Edition 2017) through the National Agency for Research (ANR)
• Leducq Foundation (Research Grant 17CVD01 titled "Gut Microbiome as a Target for the Treatment of Cardiometabolic Diseases").
• K Santé Society providing meals for the study.
• Nutrisens Society providing snacks for the study.