Revealing the unique signatures of healthy guts as we age

Why study gut microbes?

Microbes are the elders of life on Earth — they were here long before us, and they’ll be here long after we’re gone. They drive the biogeochemistry of the planet; without them, the composition of the atmosphere, crust, and oceans would be completely different. This extends to our own bodies — we’ve outsourced a lot of the human body’s functionality to commensal microbial ecosystems that help to keep us healthy. I started out in environmental microbial ecology, and became interested in ecosystem restoration and how ecosystem function can be degraded by disturbances or invasive species. Over time, my interests have drifted to the microbial ecology of the human body in both health and disease. Ecosystem function and ecosystem health can be difficult to define and quantify in environmental systems. However, there is a single elegant question you can ask regarding how well the ecosystem of the body is functioning, which is simply: how do you feel?

In what ways are healthily ageing microbiomes unique?

Our previous work confirmed that there are two clearly defined signatures in the guts of elderly people¹. We examined multiple large, multi-omics dataset that included blood metabolites, blood proteins, clinical chemistries, health and lifestyle data, and stool microbiome samples from people sampled across the lifespan. The healthy elderly had a markedly different microbiome composition from those who were unwell, and their microbiomes were quite distinct from younger people. In the healthy elderly, the core microbes that we all share were declining in their dominance, while sub-dominant taxa were rising in prominence. This essentially gave each individual a more unique gut microbiome ‘fingerprint’— healthy elderly were drifting apart from one another in terms of their microbiome composition. In elderly people who had health problems, this ‘uniqueness’ signal disappeared. We found that uniqueness was directly associated with a specific group of microbially-derived blood metabolites, which were more prevalent in healthy elderly. A few of these blood metabolites have been shown to extend lifespan in non-human animal models. People with a more unique signature were more likely to survive at the four-year follow-up mark than those without.

So it may not be helpful to ‘turn back the clock’?

Precisely. We would suggest caution in giving an older person a younger person’s microbiome. The younger person’s microbiome may contain the wrong commensal bacteria for an older person to age well. For example, there are many species of the genus Bacteroides, which are particularly common in the guts of younger people. This genus specializes in breaking down complex carbohydrates — some break down dietary carbohydrates, others degrade host-derived carbohydrates like mucus glycans, and some can eat both. As we age, our appetite declines, and so the primary source of nutrition for Bacteroides skews more towards host mucus — these microbes may pivot more towards degrading that mucus layer, leading to inflammation and ill health. Also, as we age, the intestinal stem cells in the epithelium start to slow down and stop dividing so readily. Thus, older people produce fewer goblet cells that generate mucus, and this can compound inflammation associated with the thinning of the mucus layer. If this hypothesis is true, giving an older person a younger person’s Bacteroides-dominated microbiome could do more harm than good.

Please outline your proposed GGGH project.

Mechanistically, it’s not clear what is causing this divergence of increasingly unique microbiomes in healthy elderly people. Is this correlation or causation, in terms of healthy ageing? Are microbially derived blood metabolites contributing to healthy ageing? Almost half the metabolites in the blood are affected by variation in the microbiome². These bioactive molecules are circulating throughout the body and interacting with our organ systems, influencing physiology and gene expression. In our first ageing study, we identified a number of microbially-derived metabolites that were closely linked with uniqueness. The strongest hit was a metabolite called phenylacetylglutamine, which was highly elevated in the bloodstreams of people with a high uniqueness score. Indeed, this molecule is already patented as a biomarker for healthy ageing, and is found not just in centenarians, but in their children too, suggesting a common genetic or lifestyle link. We plan to augment our existing data set by doing metagenomic sequencing. We’ll dive into the microbial genes and see if we can determine which microbial functions are associated with the uniqueness signature in healthy elderly people.

What metagenomic links might you find?

We ran an initial analysis on 30 people from our database with metagenomic data. We found that genes involved in utilizing or consuming indoles (a group of microbially-derived blood metabolites) were declining with age, while genes involved in synthesizing indoles were increasing. Similarly, we saw higher levels of indole in the blood of individuals with more unique microbiomes. Together, these results suggest a mechanistic link between microbial indole production and the availability of indole in the host bloodstream. We will examine all the different blood metabolites in healthily ageing people in this way. We will also examine other microbial metabolites produced in the gut, like short-chain fatty acids; these have a strong impact on a person’s health but aren’t present in high levels in the blood. Examination of the metagenome will allow us to identify the genes that are processing or synthesizing these metabolites, and how this might be impacting gut health.

How will computational modelling contribute?

The biggest technical challenge here is determining causality, and this is particularly difficult in humans. Work in non-human animal models rarely translates directly to humans, especially with a highly complex and dynamic ecosystem like the gut. We developed a community-scale metabolic modelling platform called MICOM that integrates genome-scale metabolic models for different taxa into a community-scale model that can be personalized to an individual using dietary and microbiome compositional data. Essentially, MICOM is a digital twin of a person’s gut: a symbolic artificial intelligence (AI) platform, which leverages detailed knowledge about microbial metabolism to make causal, mechanistic predictions of how an individual’s gut microbiome transforms dietary and host substrates into a vast array of microbial metabolites in the gut.

What will the practical applications of your study be?

We will use MICOM to simulate the metabolic functionality of the gut ecosystem. We can then examine the plasticity of the resulting microbial behaviours by manipulating dietary inputs, including prebiotics and probiotics. Hopefully, we can pinpoint simple, non-pharmaceutical interventions that augment metabolites associated with healthier ageing, or deplete those that could be harming health in older age. Building and improving AI models like MICOM will enable future human intervention trials where we can test personalized prebiotic, probiotic, and dietary interventions designed to prevent or even reverse ageing-related chronic diseases.

Biography

Sean Gibbons is an associate professor at the Institute for Systems Biology (ISB), an affiliate associate professor in the Departments of Bioengineering and Genome Sciences at the University of Washington, and a data science fellow at the eScience Institute, in Seattle, USA. His lab studies the human microbiome and its influence on health and disease, spanning the fields of ecology, evolution, microbiology, biomedicine, and computational systems biology. The lab has recently developed community-scale metabolic modelling tools for engineering the metabolic outputs of the gut microbiome, which has potential applications to precision nutrition and personalized probiotics.