Published online 26 October 2011 | Nature | doi:10.1038/news.2011.614


Friendly bacteria move in mysterious ways

Probiotic yoghurts have only a small effect on gut bacteria.

yoghurtProbiotic yoghurts seem to have little effect on the population of bacteria in the human gut.Van Osaka/Photononstop/

Many yoghurts are loaded with live bacteria, and labelled with claims that consuming these microorganisms can be good for your health. But a study published today shows that such yoghurts have only subtle effects on the bacteria already in the gut and do not replace them.

Nathan McNulty, a microbiologist at Washington University in St Louis, Missouri, recruited seven pairs of identical twins, and asked one in each pair to eat twice-daily servings of a popular yoghurt brand containing five strains of bacteria.

By sequencing bacterial DNA in the twins' stool samples, the team showed that the yoghurt microbes neither took up residence in the volunteers' guts, nor affected the make-up of the local bacterial communities.

Jeffrey Gordon, the microbiologist at Washington University who led the study, was not surprised. "We were only giving several billion bacterial cells in total to the twins, who harbour tens of trillions of gut microbes in their intestines," he says.

Murine model

McNulty also fed the five bacterial strains from the yoghurt to 'gnotobiotic' mice — animals raised so that the only microorganisms that their guts contain are 15 species found in humans.

As with the twins, the yoghurt bacteria did not change the composition of the rodents' resident communities. However, the activity of genes that allow the native bacteria to break down carbohydrates did increase. One of the five yoghurt strains — Bifidobacterium animalis lactis — also showed a similar boost in its ability to metabolize carbohydrates.

The study, which was funded by the US National Institutes of Health and Danone Research, the research arm of the food company that makes some probiotic yoghurts, was published in Science Translational Medicine1.

Companies that sell foods with added ingredients that are intended to boost health or prevent illness are under increasing pressure to substantiate the claims about their products. The pressure was increased earlier this year, when the European Food Safety Agency criticized many products, following an extensive three-year review.

The health claims are hard to test, says Gordon, because there is so much variety in the bacteria in the yoghurts and in the people who consume them, but gnotobiotic mouse models will be vital for such studies. Using the mice, he can examine the effect of probiotic foods under tightly controlled conditions, with "defined communities where all the actors and genes are known". The mouse models "provide a foundation for critically evaluating the claims from manufacturers of functional foods and probiotics", he says.

More twins needed

It's too early to be drawing conclusions, say other researchers.

Dusko Ehrlich, a microbiologist at the French National Institute for Agricultural Research (INRA), thinks that the team did not look at enough twins, or sequence their bacterial genes at enough depth. "We should reserve the judgement on the effects of probiotics in humans until broader and deeper studies are carried out," he says.

Jens Walter, a microbiologist at the University of Nebraska-Lincoln calls the work an "impressive technical achievement". However, he cautions that there are limits to studying mice with human gut bacteria, because different species have their own specifically evolved sets.

McNulty acknowledges the problem. But, he says, "we're always looking to refine these mouse models to be more like the human context. The fact that we see shared responses in mice and humans is good evidence that we're doing something right."

Gordon now wants to find out exactly how the ingested microbes influence the entrenched ones. "These are very dynamic systems with a lot of cross-talk," he says. "We don't know the nature of the conversation that's occurring." 

  • References

    1. McNulty, N. P. et al. Sci. Trans. Med. 3, 106ra106 (2011). | Article |
Commenting is now closed.