Getting to know the gut microbiome

The microbiome is taking over the world. Or, more precisely, we humans are finally cottoning to the power wielded over our health by communities of microbes living in the different niches of our bodies. Over the past few years, researchers have linked the human microbiome — particularly the community residing in the gut — to a multitude of diseases and conditions, including obesity, Crohn’s disease, depression, multiple sclerosis and Parkinson’s disease. And treatments such as fecal transplants (yes, they are just what they sound like) against life-threatening infections by the bacterium Clostridium difficile hint at the power of manipulating these microbial communities to health-promoting ends.

It’s fair to say that no one has done as much to develop the DNA sequencing and computational tools for studying the composition and function of the microbiome as genomicist Rob Knight of the University of California, San Diego. In “The Microbiome and Human Biology,” published in the Annual Review of Genomics and Human Genetics, he and his colleagues describe the technological advances that have brought researchers so far, as well as current bottlenecks and future prospects for the field.

This interview has been edited for length and clarity.

You’re learning that human microbial communities can be enormously different. Was that a surprise?

It was and it wasn’t. Antonie van Leeuwenhoek showed in the 1670s to 1680s that the different parts of the body had different microbes, so in a sense the Human Microbiome Project [a $115-million National Institutes of Health initiative to characterize the microbiomes of healthy humans] was scooped by about 300 years. But what’s amazing is that the differences between, say, a sample of dirt and a sample of water are comparable to the difference between your mouth and your gut. This remarkable level of heterogeneity in different sites of the human body was really unexpected.

What about the difference in microbial communities between individuals?

That was the other big surprise. The initial assumption of the Human Microbiome Project was that it would turn out kind of like the Human Genome Project, where everyone’s basically the same. But it wasn't like that at all. Different people have totally different microbiomes. It’s been very surprising to find that there isn't really a core microbiome but, rather, these very highly individualized microbiomes in different people.

Are there still technological challenges to determining the composition of the microbiome?

There is some fundamental scientific interest in probing the composition, but what's really exciting is using the composition as a tool to reveal something about the physical and biochemical characteristics of the microbiome, and to understand the mechanisms behind how these characteristics change.

For example, our recent work with Sarkis Mazmanian [at the California Institute of Technology] looked at the role of the gut microbiome in Parkinson’s disease, which you might not expect to be linked. What’s interesting is, we can work out specifically how different kinds of microbes are stimulating the immune system, and how that interacts with host genetics, and the rest of the microbiome, to either increase or decrease susceptibility to Parkinson’s.

Scientists are starting to figure out ways to manipulate the microbiome to benefit health. What do they need to know to expand that capability to a wider range of health conditions?

We know that a whole range of different interventions, such as fecal transplants, as well as dietary changes or taking antibiotics, can cause statistically significant changes in the microbiome. We also know there's a whole lot of disease states that link to differences in the microbiome. What we need to know is how to change the microbiome in a particular, defined way. We also need to know whether the direction we are changing it in is good or bad.

Essentially, we need to integrate data from a huge number of different clinical trials, and different studies that follow people over time, so that you can compare the effects. Then we'll have a much better sense of whether an intervention is changing the microbiome in a good or a bad direction.

What do you see as the bottleneck to doing that?

Everything's being done with different methods. That can often outweigh the biological effects that you're looking for. It’s really a question of getting people to agree to use standard approaches. That’s hard, because a particular approach might be better for studying one site of the body versus another, or one disease versus another, or better for studying, say, the gut versus soil.

How do researchers search for causal connections between the microbiome and various diseases?

We have some very nice tools, especially transferring fecal samples — or microbes or molecules isolated from fecal samples — into germ-free mice, and being able to transfer the physiological effects from individual humans across the species boundary. That’s been very exciting; I've been involved in such work on obesity, malnutrition, Parkinson’s disease, multiple sclerosis. Others have done it for autism, depression, diabetes, and a range of other diseases.

But it’s important to keep in mind that although transplanting microbes from fecal samples into germ-free mice is useful for identifying how exactly the microbiome affects a particular system in the body, it doesn’t necessarily tell you what component of the system is the key one to tinker with in treating human health conditions. A great example of this is the hormone leptin, which is an important component of the circuit that regulates weight, though it’s not the controlling factor in most humans unless you are leptin-deficient. Pretty much any fat mouse can be slimmed down with a leptin injection, but that's not true for fat humans. It’s always important to remember that humans are not mice, and translating animal results to humans is hard.

If you could imagine yourself in a sci-fi world with super-advanced technology, what really cool things could microbiome researchers do that they can’t do now?

Microbiome research at the moment is where photography was 100 years ago: If you wanted to be a photographer you could buy a camera, but you had to make all the plates and mix all the reagents yourself and build your own darkroom, set up the developing stuff, and so on. Compare that to the camera on your cell phone, which you can use for all sorts of applications. Maybe you snap a photo of your shopping list and take it to the supermarket with you — there’s no way in hell you would do that if you had to wait to get it developed.

I want the equivalent of the cell phone camera for microbial genomes and microbial metabolism — something instantly accessible, with a user interface that also allows you to make sense of the data and use the data. So that you can just press “play” for your microbiome.

What are the big unanswered questions about the human gut microbiome, and the next research frontiers?

Question one: Is there an optimal gut microbiome for everyone, or is it highly personalized? Can we discover the factors that lead to that personalization to shift your microbiome to an optimal state?

Question two: How much can we predict from your microbiome today what will happen to you, not just tomorrow, but ten, fifteen, twenty years down the track?

And question three: How much do the radical differences in the microbiomes of people around the world stem from their lifestyles, and how much do those differences lead to different susceptibility to disease, and different ability to metabolize various foods and drugs? And can we come up with a public health program for modifying the microbiome that’s as effective as were the whole-population-level nutritional supplementation programs instituted in the US during World War II?