Metabolite-based dietary supplementation in human type 1 diabetes is associated with microbiota and immune modulation

Abstract

Background: Short-chain fatty acids (SCFAs) produced by the gut microbiota have beneficial anti-inflammatory and gut homeostasis effects and prevent type 1 diabetes (T1D) in mice. Reduced SCFA production indicates a loss of beneficial bacteria, commonly associated with chronic autoimmune and inflammatory diseases, including T1D and type 2 diabetes. Here, we addressed whether a metabolite-based dietary supplement has an impact on humans with T1D. We conducted a single-arm pilot-and-feasibility trial with high-amylose maize-resistant starch modified with acetate and butyrate (HAMSAB) to assess safety, while monitoring changes in the gut microbiota in alignment with modulation of the immune system status.

Results: HAMSAB supplement was administered for 6 weeks with follow-up at 12 weeks in adults with long-standing T1D. Increased concentrations of SCFA acetate, propionate, and butyrate in stools and plasma were in concert with a shift in the composition and function of the gut microbiota. While glucose control and insulin requirements did not change, subjects with the highest SCFA concentrations exhibited the best glycemic control. Bifidobacterium longum, Bifidobacterium adolescentis, and vitamin B7 production correlated with lower HbA1c and basal insulin requirements. Circulating B and T cells developed a more regulatory phenotype post-intervention.

Conclusion: Changes in gut microbiota composition, function, and immune profile following 6 weeks of HAMSAB supplementation were associated with increased SCFAs in stools and plasma. The persistence of these effects suggests that targeting dietary SCFAs may be a mechanism to alter immune profiles, promote immune tolerance, and improve glycemic control for the treatment of T1D.

Trial registration: ACTRN12618001391268. Registered 20 August 2018, https://www.anzctr.org.au/Trial/Registration/TrialReview.aspx?id=375792 Video Abstract.

Keywords: Autoimmunity; Dietary-metabolites; Immune regulation; Microbiota; SCFAs; Type 1 diabetes.

Fig. 1

Experimental study in patients with T1D. Schematic diagram illustrating the study design and selection procedure for the individuals enrolled in the study

Fig. 2

Increased concentration of short-chain fatty acids in stool and plasma following HAMSAB supplementation. A Acetate, propionate, and butyrate concentrations in stool (mM) and B plasma (μM). Overall significance determined by GEE and pairwise differences between timepoints by estimated marginal means and include a Tukey adjustment for multiple corrections. Colors indicate individual subjects. Box plots show mean and upper and lower quartile ranges

Fig. 3

Composition and function of the gut microbiome is altered following HAMSAB supplementation. A Multivariate sPLS-DA comparing microbial species present at each timepoint. Significance determined by PERMANOVA. Loadings shown in Fig S1. B Alpha diversity measured by inverse Simpson index. Overall significance determined by GEE and pairwise differences between timepoints by estimated marginal means. Colors indicate individual participants. Box plots show mean and upper and lower quartile ranges. C Mean log foldchange in relative abundance from baseline at W6 and W12, grouped by taxonomic classification. Asterix represents GEE significance of changes in abundance from baseline. Error bars represent standard deviation. # adjusted P < 0.1, *adjusted P < 0.1–0.05, **adjusted P < 0.01, ***adjusted P < 0.001. D Multivariate sPLS-DA comparing microbial pathways present at each timepoint and plot loadings indicating the contribution of each bacterial function to the variance. Color corresponds to the timepoint

Fig. 4

HAMSAB supplementation is accompanied by modulation of the immune system at W6 and W12 follow-up. A Multivariate PLS-DA comparing the proportions of major immune populations assessed by mass cytometry within total live cells. Significance determined by PERMANOVA. B PLS-DA plot loadings indicating the contribution of each immune population. C Proportions of total CD19⁺ B cells and IgD⁺CD27^- naïve B cells within live cells. D IgD⁺IgM^hiCD27⁺ MZ B cells expressing CD86 (mean geometric signal intensity, MSI). E CD3⁺ T cell % within live cells. F CTLA4 expression (MSI) on granzyme B⁺ perforin⁺ (Grzb⁺Perf⁺) Tconv CD4⁺ T cells and Grzb⁺Perf⁺ CD8⁺ T cells. G TIGIT expression (MSI) on TEMRA Tregs, CM Tconv CD4⁺ T cells and % TIGIT⁺CD45RO⁺ Tconv within live cells. Colored dots and lines represent each subject. Box plots show mean and upper and lower quartile ranges. Significance determined by GEE. Adjusted P values are (6W vs W0) or (12W vs W0). Gating strategy shown in Fig. S2

Fig. 5

Circulating pro-inflammatory mediators are reduced at W12 in subjects following HAMSAB supplementation. A Serum IL-8, MIP1a, and bFGF concentrations detected by multiplex assay. Overall significance determined by GEE and pairwise differences between timepoints by estimated marginal means. Box plots show mean and upper and lower quartile ranges. B Hierarchical clustering of genes from fatty acid metabolism KEGG pathway gene set and from oxidative phosphorylation KEGG pathway gene set at baseline and 6 weeks of HAMSAB supplementation (FDR = 0.015 and 0.002, respectively). Columns represent individual subjects and rows represent individual genes in the pathway

Fig. 6

Increased SCFAs correlated with changes in glycemic control, commensal microbiota, and immune cell changes. A Heatmap of Pearson r values between relative abundance of SCFAs and clinical data. *Adjusted P < 0.05. **Adjusted P < 0.01. Stool and plasma short-chain fatty acids are hierarchically clustered based on Bray-Curtis dissimilarity. B Pearson r values at each timepoint between plasma butyrate, HbA1c, and daily basal insulin. Grey shading represents 95% confidence interval. C Heatmap of regression coefficients determined by GEEGLM between bacterial taxa and pathways with stool and plasma SCFAs and glycemic markers across all three timepoints. Bacterial pathways and taxa are hierarchically clustered based on Bray-Curtis dissimilarity. D Heatmap of significant regression coefficients across all three timepoints determined by GEEGLM between bacterial taxa that significantly changed across time (adj P < 0.05) and/or those correlated with SCFA and glycemic markers in (C) and significantly altered immune subsets (adj P < 0.05). *Adjusted P < 0.05, **adjusted P < 0.01, ***adjusted P < 0.001