# Translating in vitro gut microbiota models to human context: compositional correlations under dietary fiber intervention

## TL;DR

This study shows that in vitro gut microbiota models can predict how human microbiota respond to dietary fiber interventions.

## Contribution

The study demonstrates that in vitro models can correlate with in vivo responses to dietary fiber in human gut microbiota.

## Key findings

- A significant correlation between in vitro and in vivo microbiota composition was observed after 8- and 12-week interventions.
- Microbial taxa responding to the fiber intervention overlapped significantly between in vitro and in vivo settings.
- In vitro models can help pre-select donors whose microbiota are likely to respond to specific interventions.

## Abstract

Large interindividual variation in human gut microbiota composition and its response to interventions limits the development of novel microbiota-targeted supplements. In vitro models reflecting this interindividual variation and predicting individual in vitro microbiota responses would allow for the assessment of the potential efficacy of such interventions.

Here, we investigated whether in vitro microbiota modulation by a dietary fiber mixture is translatable to in vivo microbiota outcomes. A 12-week double blind, randomized, placebo-controlled, crossover study with a dietary fiber mixture of acacia gum (AG) and carrot powder was performed in healthy volunteers (N = 54, 45–70 years, BMI 27.3 ± 1.4 kg/m2). The in vitro platform utilized fecal samples from the same individuals who participated in the in vivo study.

A significant effect on microbiota composition was shown in vivo, although with strong individual variation. The fiber intervention was mimicked in vitro by exposing each individuals’ baseline microbiota to the same dietary fiber as used for the 12-weeks in vivo intervention. A significant correlation was shown between the in vitro and human fecal microbiota composition after 8- and 12-weeks intervention (p = 0.003 and p = 0.0107, respectively). Microbial taxa responding to the intervention in vitro and in vivo also showed clear overlap (p = 0.002).

These results demonstrate that in vitro models may enable pre-study selection of donors whose microbiotas respond to a specific intervention.