Gut Microbiome Links to Age-Related Traits and ApoM Protein
Federica Grosso from the Institute for Genetic and Biomedical Research (IRGB) of the National Research Council (CNR) in Monserrato, Italy, describes a research paper she co-authored that was published in Volume 17, Issue 8, titled “Causal relationships between gut microbiome and hundreds of age-related traits: evidence of a replicable effect on ApoM protein levels.”
DOI - https://doi.org/10.18632/aging.206293
Transcript
Hi everyone. My name is Federica Grosso, and I’m a research fellow at the Institute for Genetic and Biomedical Research at the National Research Council of Italy. Today, I’m going to present the results of our paper recently published in the journal Aging. The full article is available open access from the Aging journal website. This link here.
This work was carried out together with Dr. Daniela Zanetti and Dr. Serena Sanna researchers at the same institute. So what was the aim of our study? We know that the gut microbiome, which is the collection of microorganisms living in our digestive system, plays a key role in the immune function and metabolic health. As we age, this microbial community changes often leading to imbalances that have been associated to inflammation and chronic diseases. However, it’s yet unknown which of these associations are the consequence of microbiome imbalance, or if they are the cause of microbiome imbalance.
To investigate if these age-related changes in the microbiome are the cause of age-related diseases and conditions, so we applied the genetic epidemiology approach called Mendelian randomization. We use as exposure public available genome-wide association summary statistics of gut microbiome features from the Dutch Microbiome Project. And as outcome, do a summary statistics of common aging traits and inflammatory and cardiometabolic proteins from the UK Biobank.
But how does Mendelian randomization work? MR uses genetic variants, which are fixed at birth and robustly associated with exposure of interest. This helps us to avoid a typical confounding that can distort observational studies and provides a cost-effective alternative to randomized control trials. We call these genetic variants, the instrumental variables or IVs.
There are tricky assumptions that must hold in order to apply this technique. The first one is the relevance exposure, which states that IV is directly associated with exposure. The second one is the independence assumption, which states that IV is not associated with the outcome due to confounding pathways.
And finally, we have exclusion restriction, which states that IV affects the outcome only through the exposure, not via other pathways. We use a technique that is called two-sample MR, which leverages GWAS summary statistics for both exposure and outcomes.
This is the workflow of our analysis. We performed the careful quality control of the GWAS and selected 37 gut microbiome traits reaching a GWAS significant value threshold of 5 x 10 to the power of -8. We then use the P value of 5 x 10 to the power of -6 to select instrumental variables. We then run MR analysis along with a comprehensive set of sensitivity checks, including bidirectional MR. Finally, we replicated the most significant relationships in independent datasets.
We had 91 significant causal relationships. For example, we found that higher levels of order of Coriobacteriales and family of Coriobacteriaceae, which is of the same order, are causally linked with an increased risk of age-related macular degeneration.
But one of the most intriguing findings came from the analysis of UK Biobank proteins. Indeed, we observed a causal relationship between the purine nucleotide degradation second pathway and apolipoprotein M levels. This result was independently replicated in another cohort. This finding is important because ApoM enables HDL to promote cholesterol efflux and reduce inflammation, and we know that this pathway allows microbes to break down purines into urate, which is a risk factor for cardiovascular health.
In our results, the increased inactivity of the pathway causes lower level of ApoM, which suggests a possible increase in cardiovascular risk increasing the urate. Another interesting results is the identification of causal relationships between the lactose and the lactose degradation first pathway and some circulating protein levels involved in both cardiovascular and inflammatory circuits. We were intrigued by these results and we decided to follow it up further.
So using MR and existing knowledge of host interaction function of fecal and bacterium transmitted strains, we confirmed that higher abundance of lactose and lactose degradation pathway is driven by the higher GalNAc utilization in blood type individuals. And this tool increases or decreases the levels of these proteins.
Based on these results, we speculate that the difference in disease risk among blood type groups for age-related conditions on which these proteins are potentially involved may be different due to a modulation of gut microbiome on host protein levels.
In particular, some studies demonstrated that TREM2 is protective for arteriosclerosis and the increase of the presence of some other proteins like SERPINA3 and C7 led to an increase of cardiac mortality and coronary artery disease respectively.
So we had some interesting results and we validated our significant findings using independent US data set. The crucial step that is often overlooked in the MR field. Although more research is needed to fully understand this biological results, our findings suggest that targeting the gut microbiota could become a promising strategy to delay or reduce age-related inflammation or diseases. Future interventions might include dietary changes, probiotics, or other microbiome-based therapies.
Finally, I’d like to thank my co-authors, Daniela Zanetti and Serena Sanna for their scientific oversight and collaboration. Davide Murrau for his technical support as well as Valeria Lo Faro and Daria Zhernakova for the critical reading and the helpful advice. We also gratefully acknowledge the fundings from CNR of Dr. Serena Sanna, and the Marie Curie Fellowship of Dr. Daniela Zanetti. Thank you.
