A probiotic approach identifies a Treg-centred immunoregulation via modulation of gut microbiota metabolites in people with multiple sclerosis and healthy individuals

Abstract

Background: Modulation of the gut microbiota composition has been suggested as a potential disease modifying therapy in immune-mediated diseases such as multiple sclerosis (MS). However, a conclusive mechanism linking gut microbiota modulation to peripheral immune responses has remained elusive so far.

Methods: In this exploratory cohort study, people with MS (pwMS) and healthy controls (HC) supplemented a lactobacilli-rich probiotic for two or six weeks and were additionally investigated six weeks after the last intake. Immune cell phenotyping was performed in blood samples, complemented by mRNA expression analysis, serum cytokine measurements, and Treg suppression assays. Besides gut microbiota composition analysis, metabolite production was investigated in stool and serum. Links between metabolites and peripheral immune system were investigated in in vitro T cell differentiation assays.

Findings: In peripheral blood, Treg cells increased in both groups, while Th1 cells were significantly reduced in pwMS. This promotion of a regulatory immunophenotype was complemented by increased concentrations of IL-10 in serum and higher expression of IL10 and CTLA4. Functional assays revealed an enhanced suppressive capacity of Treg cells due to the probiotic intervention. The tryptophan metabolite indole-3-acetate (IAA) increased in stool and serum samples of pwMS during the probiotic intake. In vitro, IAA specifically enhanced the formation of IL-10 secreting T cells together with CYP1a1 expression. This effect was blocked by addition of an aryl hydrocarbon receptor (AHR) inhibitor.

Interpretation: A lactobacilli-enriched probiotic promotes a regulatory immunophenotype in pwMS, probably by enhancing AHR agonists in the gut. It may be of interest as add-on therapy in immune-mediated diseases such as MS.

Funding: This study has in part been funded by Novartis Pharma GmbH and BMBF grant no. 01EJ2202B.

Keywords: AHR signalling; Enhanced Treg suppressive capacity; Immunomodulation; Indole-3-actetic acid; Multiple sclerosis; Probiotic supplementation.

Fig. 1

Study design and baseline characteristics. This exploratory study investigated immunomodulatory effects of a lactobacilli-containing oral probiotic (Vivomixx®) in a per-protocol study population consisting of healthy controls (HC, n = 41) and people with MS (pwMS, n = 28). All study participants supplemented the probiotic for at least two weeks, a subgroup of 15 pwMS and 17 HC continued the probiotic intake for up to six weeks. PwMS supplemented the probiotic to their existing immunomodulatory therapy. Parameters were analysed before probiotic intervention (baseline) and after two or six weeks of daily probiotic intake and 6 weeks after termination of probiotic supplementation.

Fig. 2

Probiotic supplementation shifts the peripheral T cell composition towards an anti-inflammatory phenotype in HC and pwMS. PBMCs from HC and pwMS were analysed for their respective marker expression via flow cytometry before (baseline), after two and six weeks of probiotic supplementation and six weeks after termination of probiotic intake (WO). The graphs show relative changes of the analysed cell population normalised to each individual's baseline. Black lines represent individual study participants. Orange lines indicate the median change. (A, B) Two weeks of probiotic intake did not significantly change IL-17A+ Th17 cells in HC (A) and pwMS (B). (C, D) IFNγ+ Th1 cells revealed no effect in HC (C) but a decrease in pwMS (D) after two weeks. (E–G) FoxP3+ Treg cells increased in HC (E) and pwMS (F). (G) Representative dot plots of FoxP3+ Treg cells demonstrate the increase after two weeks of probiotic intake (right) compared to baseline (left). (H, I) Six weeks of probiotic intake did not change significantly Th1 cell frequencies in HC (H) but decreased Th1 cells in pwMS (I). (J, K) Treg cell frequencies were significantly increased in HC (J) and pwMS (K) after six weeks. (L–O) Th1 and Treg cell frequencies were analysed 6 weeks after the last probiotic intake (washout, WO). Relative frequencies of Th1 cells reduced in HC (L) and pwMS (M), whereas Treg cell frequencies remained unchanged in HC (N) but dropped in pwMS (O) compared to baseline frequencies. Significance for changes over time were calculated using unnormalised cell frequencies with Wilcoxon matched-pairs signed rank test. (A) n = 41, p = 0.672. (B) n = 28, p = 0.333. (C) n = 40, p = 0.952. (D) n = 28, ∗p = 0.029. (E) n = 37, ∗∗p = 0.005. (F) n = 28, ∗p = 0.036. (H) n = 17, p = 0.963. (I) n = 15, ∗p = 0.0302. (J) n = 17, ∗p = 0.0295. (K) n = 15, ∗p = 0.0129. (L) n = 17, p = 0.404. (M) n = 15, p = 0.064. (N) n = 17, p = 0.382. (O) n = 15, p = 0.629.

Fig. 3

Probiotic supplementation increases IL-10 production and the suppressive capacity of Treg cells. (A–C) mRNA expression analysis in PBMCs of HC and pwMS was analysed at baseline and after two weeks of probiotic intake. (A) IL10 gene expression significantly increased in pwMS after two weeks of probiotic supplementation but not in HC. (B) CTLA4 gene expression significantly increased in pwMS after two weeks of probiotic supplementation but not in HC. (C) IFNG gene expression remained unchanged in pwMS and HC. Graphs show individuals as points and mean ± 95% CI. (D, E) The concentration of IL-10 was analysed by ELISA in serum samples of HC (D) and pwMS (E) at baseline and after two weeks of probiotic intake. The IL-10 concentration increased in pwMS but not in HC. Graphs show relative changes normalised to each individuals' baseline. Black lines represent individual study participants. The orange line indicates the mean change. (F) Representative histograms of the in vitro suppression assay show the proliferation of CD4+ PBMCs isolated from HC before (left) and after 2 weeks of probiotic intake (right). (G) Treg cell suppressive capacity in HC increased after two weeks of probiotic supplementation. Significance for changes over time was calculated using unnormalised values with Wilcoxon matched-pairs signed rank test in A–G. (A) HC: n = 40, p = 0.727; pwMS: n = 25, ∗p = 0.019. (B) HC: n = 40, p = 0.332; pwMS: n = 25, ∗∗p = 0.004. (C) HC: n = 40, p = 0.793; pwMS: n = 25, p = 0.353. (D) n = 40, p = 0.179. (E) n = 28, ∗∗p = 0.002. (G) n = 11, ∗p = 0.0186.

Fig. 4

Probiotic bacterial strains efficiently colonise the gut microbiome of HC and pwMS. 16S rRNA sequencing was performed in stool samples from HC and pwMS before (baseline) and after two weeks of probiotic supplementation (HC n = 27 per timepoint; pwMS n = 22 per timepoint). (A) Alpha diversity of the gut microbiome in pwMS (green) and HC (blue) revealed a higher microbiota richness in HC compared to pwMS but no effect due to probiotic supplementation. (B) Beta diversity as presented by Bray–Curtis permutational multivariate analysis of variance differed between HC and pwMS, but it was not affected by probiotic intake. (C) Differential abundance analysis with the Linear Model for Differential Abundance Analysis of High-dimensional Compositional Data (LinDA) revealed a differential abundance of bacterial genera between HC and pwMS at baseline. (D) Differential abundance analysis with LinDA confirmed a successful colonisation of all probiotic bacteria in the gut microbiota after two weeks of probiotic supplementation. (E) The enrichment of all probiotic bacteria could be confirmed in both study groups (green: HC, blue: pwMS; left: 2 weeks of probiotic intake, right: baseline). Differences between cohorts were analysed using a Wilcoxon rank-sum test at baseline and after two weeks of treatment. p-Values were adjusted for multiple comparisons using the Benjamini–Hochberg method.

Fig. 5

Probiotic supplementation increases indole-3-acetic acid (IAA) in stool and serum and IAA enhances IL-10 production in T cells in vitro. (A) Heat-map of microbial metabolites analysed in stool samples of pwMS before probiotic supplementation (baseline) and after two weeks by high performance liquid-chromatography–mass spectrometry. (B, C) Concentrations of IAA in (B) stool samples and (C) serum samples of pwMS before probiotic intake (baseline) and after two weeks. Black lines represent individual pwMS. The graphs show relative changes normalised to each individual's baseline. Wilcoxon matched-pairs signed rank test B: n = 22, p = 0.0460; C: n = 7, p = 0.1094. (D) Naïve CD4+ T cells from HC were cultured under co-stimulatory anti-CD3/anti-CD28 conditions in the presence or absence of different concentrations of IAA. IAA increased CYP1a1 gene expression in in vitro cultured T cells (n = 5 per group; Kruskal–Wallis test with post-hoc Dunn's test and Bonferroni adjustment for multiple comparisons, p = 0.0277; p = 0.0037). (E–G) Naïve CD4+ T cells from HC were cultured under co-stimulatory anti-CD3/anti-CD28 conditions in the presence or absence of 300 μM IAA with or without 5 μM of the AHR inhibitor CH-223191. (E) The IAA induced increase in CYP1a1 gene expression was prevented by simultaneous addition of CH-223191 (n = 5 per group). (F) Flow cytometry analysis of IL-10+ cells among CD4+ T cells revealed a relative increase under IAA and no effect upon additional treatment with CH-223191 (n = 8 per group; Two-way ANOVA with Šidák's Multiple comparisons test, factor 1: IAA concentration; factor 2: CH-223191 addition). (G) IL-10 concentrations were analysed in cell culture supernatants by ELISA and revealed a significant increase in IAA treated T cells that was blocked by the simultaneous addition of CH-223191. (n = 8 per group; Two-way ANOVA with Šidák's Multiple comparisons test; factor 1: IAA concentration; factor 2: CH-223191 addition).